Abstract

The processes by which direct-care workers identify psychological disorder and need for treatment in adults with mental retardation was examined. A series of vignettes—varying in problem type, client's mental retardation level, and client's gender—was administered to 144 direct-care workers from four types of work settings. Results indicate that although the workers demonstrated an overall ability to accurately identify psychopathology, problem type, and client's mental retardation level were significantly related to workers' judgments regarding the presence of psychological disorder and need for treatment. Worker's setting and client's gender were not significant predictors. Implications for staff training and for increasing the accessibility of mental health services are considered.

Editor in charge: Steven J. Taylor

There is considerable evidence that individuals with mental retardation do not receive mental health services proportionate to their need (Nezu & Nezu, 1994; Reiss, 1990, 1982; Reiss, Levitan, & McNally, 1982). A primary reason appears to be a robust minimizing bias (Lopez, 1989), whereby the psychological problems and treatment needs of persons with mental retardation are underrecognized (Alford & Locke, 1984; Reiss, Levitan, & Szyszko, 1982; Reiss & Szyszko, 1983; Spengler, Strohmer, & Prout, 1990). Research in this area has focused on the processes by which clinicians identify psychopathology and formulate treatment recommendations. However, as persons with mental retardation rarely initiate the referral process themselves (Edelstein & Glenwick, 1997; Fletcher, 1993; Szymanski, 1980), the delivery of mental health services to this population is largely dependent upon the ability of various nonmental health service providers (e.g., residence counselors and case managers) to recognize the presence of emotional disturbance and to refer for treatment (Borthwick-Duffy, 1994; Rojahn & Tassé, 1996). Borthwick-Duffy (1994) has suggested that direct-care workers, who have the highest degree of interaction with persons with mental retardation and routinely perform informal assessments, may be ill-prepared to differentiate mental health disorders from nonclinical problems in adaptive functioning (Borthwick, 1988; Rutter & Graham, 1970).

In the research conducted to date, a number of factors have been implicated in direct-care workers' ability to identify psychopathology and need for treatment in individuals with mental retardation. These include (a) the type of problem displayed; (b) the setting within which the staff member works; and (c) client variables, such as level of mental retardation and gender.

With respect to problem type, a number of investigators have pointed to a high estimation of externalizing disorders, relative to internalizing disorders, in persons with mental retardation (Borthwick-Duffy & Eyman, 1990; Edelstein & Glenwick, 1997; Jacobson, 1990a; Menolascino, Gilson, & Levitas, 1986; Sovner & Hurley, 1983). Several researchers have suggested that direct-care workers may not have the training and skills necessary to identify signs of internalizing disorders among individuals with mental retardation, whose impaired language and conceptual skills limit their ability to self-report such symptoms (Jacobson, 1990b; Reid, 1977; Sovner & Hurley, 1983). Direct-care staff members also may tend to attend to behaviors that are more troublesome in the care and management of the individual and place extra demands on the caregivers (Borthwick-Duffy & Eyman, 1990; Bruininks, Hill, & Morreau, 1988). Researchers who investigate direct-care staff members' responses to the behaviors of individuals with mental retardation have found that behaviors that are harmful to others elicit greater staff attention from residential workers than do self-injurious or unusual behaviors (Hill & Bruininks, 1984), and aggressive and bizarre behaviors are perceived by vocational counselors to be more significant than personality characteristics and anxiety (Foss & Peterson, 1981).

Concerning setting, research suggests that the subjective importance to a specific referral source (i.e., setting) of a particular maladaptive behavior may influence the likelihood that an individual will be referred for evaluation and treatment. Edelstein and Glenwick (1997), for example, found that for adults with mental retardation, reasons for referral to a mental health clinic varied as a function of the setting from which the referral originated.

With regard to client mental retardation level, a number of researchers have demonstrated that individuals with borderline to mild mental retardation are more likely to receive a psychiatric diagnosis than are those with more severe cognitive deficits (Borthwick-Duffy & Eyman, 1990; Iverson & Fox, 1989; Jacobson, 1982, 1990b; Rojahn, 1994). Furthermore, aggressive disorders have been reported more often in individuals with more severe levels of mental retardation (Eaton & Menolascino, 1982; Edelstein & Glenwick, 1997; Jacobson, 1982), whereas internalizing disorders (e.g., depression) have been reported more often in persons with borderline to mild retardation (Benson, 1985; Jacobson, 1982; Pawlarcyzk & Beckwith, 1987; Reiss, 1982). It remains unclear whether these differences are due to actual prevalence rates or differential assessment of disorder across levels of mental retardation.

Finally, although client's gender has been found to be unrelated to the overall presence of psychopathology in individuals with mental retardation (Borthwick-Duffy & Eyman, 1990; Iverson & Fox, 1989), aggressive disorders are more frequently cited as the reason for referral to mental health services among males (Benson, 1985; Reiss, 1982), whereas depression has been reported as the basis for referral more often for females (Benson, 1985; Edelstein & Glenwick, 1997; Pawlarcyzk & Beckwith, 1987; Reiss, 1982). Researchers have yet to clarify whether these differences are the result of actual normative differences in prevalence or, rather, reflect selective information-processing (i.e., gender bias) on the part of referral sources.

In the present study we systematically varied problem type, client's mental retardation level, and client's gender in a series of clinical vignettes. These case descriptions were given to equal numbers of direct-care workers in four types of settings, along with a set of questions designed to elicit staff members' judgments relating to the presence of psychopathology and need for treatment. Associations between these judgments and the aforementioned variables were examined. By presenting respondents with standardized and systematically varied vignettes, this methodology controls for client population demographics and clients' abilities to self-report symptomatology. Therefore, significant associations between the variables of interest and direct-care worker judgments may reveal the attributional styles of these staff members in identifying psychological disorder and need for treatment.

We expected that the probability of direct-care staff members identifying the presence of psychopathology and making treatment recommendations would be related to (a) the type of pathology displayed by the hypothetical client, (b) the client's mental retardation level, (c) the client's gender, and (d) the staff member's employment setting. We also expected that staff members' perceptions of (a) the environmental disruptiveness of the problem behaviors presented and (b) prognosis with and without treatment would vary as a function of the type of problem and would influence identification of psychopathology and treatment recommendations.

Method

Participants

The sample consisted of 144 direct-care workers recruited from three nonprofit social service agencies specializing in providing rehabilitative and treatment services for adults with developmental disabilities in the metropolitan New York area. These agencies were the Federation of Employment and Guidance Services, the Association for the Help of Retarded Children, and the Young Adult Institute. Services provided by these agencies include, but are not limited to, residential, vocational, day treatment, case management, and mental health programming for adults with developmental disabilities.

Direct-care workers were defined as those personnel who work in positions of direct client contact (i.e., nonsupervisory roles) in residential, vocational, case management, and day treatment programs with individuals who have mental retardation. In addition, participants were limited to employees who did not hold advanced (i.e., master's or doctoral) degrees. From each participating agency, 12 direct-care workers were selected from each of the following four types of settings: (a) residential programs, (b) vocational programs, (c) case management programs, and (d) day treatment programs. Thus, the total sample of 144 consisted of 48 workers from each of three agencies (48 × 3 = 144), or 36 workers from each of four settings (36 × 4 = 144). This represented approximately 50% of direct-care workers at the participating programs.

Settings

Residential

The residential worker group consisted of direct-care workers employed at one of three types of residential placements. These were (a) intermediate care facilities (ICFs), (b) community residences, and (c) independent residential alternatives. The ICFs are group-living settings that provide 24-hour on-site supervision, personal training, and structured activities to individuals in need of intensive support and assistance with activities of daily living. Community residences also have 24-hour on-site supervision but provide longer-term care than do the ICFs and have a lesser degree of support with regard to activities of daily living. Independent residential alternatives are residential settings where individuals with mental retardation live in supported apartments with a limited number of hours of on-site staff support, as determined by the client's degree of need.

Direct-care workers in these settings included residence counselors and habilitation workers. Residence counselors provide training and assistance with activities of daily living in the group settings of the ICFs and community residences. Residential habilitation workers provide one-on-one training in activities of daily living to individuals in independent residential alternatives.

Vocational

The vocational worker group consisted of direct-care workers employed at one of two types of vocational programs: (a) vocational rehabilitation centers (i.e., sheltered workshops) and (b) supported employment programs. Vocational rehabilitation centers are work settings where individuals with mental retardation work with full-time staff support. These programs are aimed at providing individuals with developmental disabilities the opportunity to learn about the world of work in a highly structured and supportive environment. Supported employment programs are aimed at helping individuals with developmental disabilities make the transition into competitive employment. These programs provide initial assessment and evaluation, preemployment training (e.g., resume writing, interview skills), placement in a supported work situation in the community, and on-site job coaching.

Direct-care staff in vocational rehabilitation centers include vocational counselors and floor supervisors. Vocational counselors provide one-on-one assistance with vocational skill development, whereas floor supervisors oversee groups of clients with respect to issues of work productivity and appropriate behavior for a work setting. Direct-care workers in supported employment programs include rehabilitation associates and job coaches. Rehabilitation associates focus on skill development and job readiness, whereas job coaches assist clients at job placements with issues related to job tenure.

Case management

The case management worker group consisted of case managers providing individualized services to persons with developmental disabilities in order to identify and access needed services. Services provided by direct-care workers in these settings include needs assessment and evaluation; advocacy; referral to medical, recreational, residential, and clinical services; and assistance with benefits.

Day treatment

The day treatment worker group consisted of direct-care workers employed at developmental disabilities day treatment programs. These day treatment programs provide full-time, intensive rehabilitation services to persons with developmental disabilities through the use of structured group and activities therapies.

Direct-care workers in day treatment settings included developmental specialists and paraprofessionals. Developmental specialists lead scheduled classroom activities, such as functional skills groups, and participate in the development of individualized treatment plans. Paraprofessionals assist in the running of the programs and implement specified behavior modification plans for individual clients.

Interviews with the program directors of each of the four types of settings, conducted by the first author, confirmed that these programs functioned similarly across agencies (i.e., had comparable goals and staff composition) and served a similar range of clients (i.e., individuals with borderline, mild, moderate, and severe retardation).

Forty-three percent of the participants were between the ages of 20 and 29, 34% between 30 and 39, 13% between 40 and 49, 9% between 50 and 59, and 1% above 60. The sample was predominantly female (69%). Forty percent of the participants identified their ethnicity as White, 32% as Black, 14% as Hispanic/Latino, 10% as other, and 4% as Asian. Seventy-four percent reported that their highest level of educational attainment was a college degree and 26%, a high school diploma.

Instruments

Case vignettes

Written case vignettes describing hypothetical persons with mental retardation served as the stimuli for eliciting direct-care workers' judgments. The vignettes used in this study were designed to reflect levels of three predictor variables: (a) problem type, (b) client's gender, and (c) client's mental retardation level.

With respect to problem type, four conditions were selected: (a) aggressive disorder, (b) depression, (c) psychosis, and (d) scoliosis (i.e., a physical disorder with no concomitant psychological disturbance). The categories of aggressive disorder, depression, and psychosis were chosen because of the high reporting of such difficulties among individuals with both mental retardation and psychological disorder (Benson, 1985; Edelstein & Glenwick, 1997; Reiss, 1982, 1990). The case describing the scoliosis condition was included to reduce the possibility of a response set in the completion of the questionnaires as well as to allow for examination of staff members' judgments based upon a nonmental health case presentation.

Each vignette contained a cluster of three symptoms reflecting the type of pathology presented. For example, tantrums served as a symptom of aggressive disorder and delusions as a symptom of psychosis. The sets of symptoms for the three psychopathology vignettes were selected from items in the Reiss Screen for Maladaptive Behavior (Reiss, 1994), an instrument designed for use with informants to screen for psychological disorders in persons with mental retardation.

Client's gender and mental retardation level (borderline to mild retardation and moderate to severe retardation) were communicated through the use of a sentence explicitly describing the hypothetical person as a man or woman with either borderline to mild or moderate to severe retardation.

The cluster of three items chosen to represent each of the problem types was redrafted to form a short paragraph describing a person displaying the behavioral descriptors. One paragraph describing each of the four problem types was created. For each type, there were four gender/mental retardation level combinations, but otherwise the wording of the vignettes remained the same. A total of 16 vignettes resulted, representing all possible combinations of problem type, gender, and retardation level (4 × 2 × 2). To ensure that the cases clearly and mutually exclusively depicted the problem types they had been intended to represent, we asked a group of seven mental health clinicians who specialize in the treatment of adults with mental retardation and who were not involved in other aspects of the study to validate the vignettes. One hundred percent correct identification was made by the raters for problem type, gender, and mental retardation level.

Questionnaire

A questionnaire consisting of five items was developed to assess direct-care workers' judgments concerning the vignettes. Item 1 (“Do you think this person has a psychological/emotional disorder?”) assessed their judgments regarding the presence of psychopathology. Participants were asked to respond in a forced choice manner (i.e., yes or no). Item 2 (“What, if any, treatment do you think this person needs?”) assessed direct-care workers' judgments regarding treatment needs. Participants were given a checklist of treatment choices and asked to check the option that they believed best applied. Treatment options were (a) no treatment; (b) only on-site behavior modification; (c) verbal psychotherapy (talk therapy), with or without behavior modification; (d) psychiatric medication, with or without behavior modification; and (e) verbal psychotherapy and medication, with or without behavior modification. Item 3 (“How disruptive would this person's behavior be to your program?”) assessed direct-care workers' perceptions of the degree of environmental disruptiveness of the behavior described. This question was scored on a 3-point Likert scale, with options ranging from 1 (not at all disruptive) to 3 (highly disruptive). Item 4 (“How do you think this person's behavior would be in a year or two without treatment?”) assessed direct-care workers' perceptions of prognosis without treatment. The final item assessed their perceptions of prognosis with treatment (“How do you think this person's behavior would be in a year or two with treatment?”). Items 4 and 5 were each scored on a 3-point Likert scale, with options ranging from 1 (behavior will worsen) to 3 (behavior will improve).

Vignette-questionnaire packets

Four packets of four vignettes each were formed. Each packet included one vignette describing each of the four problem types (i.e., aggressive disorder, depression, psychosis, and scoliosis). Client's gender and mental retardation level were counterbalanced within each packet so that one case of each possible gender/retardation level would appear. In addition, problem type and gender/mental retardation level combination were balanced across packets so that, for each problem type, a different gender/mental retardation level combination would occur in each packet. Vignette order was random within packets.

Procedure

Participants were recruited at staff meetings at the three agencies. Staff members agreeing to participate were asked to read the case vignettes and answer the questions based upon the assumption that the cases represented clients in their particular work setting. Each form of the questionnaire packet was given with equal frequency to participants in each of the four types of settings.

Results

Descriptive statistics regarding the frequency of identification of psychological disorder and of each treatment recommendation are presented first. Findings regarding relationships among the variables of interest follow.

Descriptive Statistics

Identification of psychological disorder

Participants correctly identified the presence or absence of psychological disorder in questionnaire Item 1 with 89% accuracy. The breakdown of the correct responses (i.e., 73% true positives; 27% true negatives) approximated the ratio of psychological disorder to nondisorder cases presented to the participants in the questionnaire packets (i.e., 3:1), whereas the errors of identification indicated a ratio of 88% false negatives to 12% false positives.

Treatment recommendations

The responses to questionnaire Item 2 are presented in Table 1. “Verbal psychotherapy, with or without behavior modification” was the most frequent response. Recommendations of “verbal therapy and medication, with or without behavior modification” and “no treatment” followed in frequency. “Psychiatric medication, with or without behavior modification” and “on-site behavior modification only” were the least frequent recommendations.

Table 1

Frequencies (ns) and Percentages of Responses Regarding Treatment Recommendation

Frequencies (ns) and Percentages of Responses Regarding Treatment Recommendation
Frequencies (ns) and Percentages of Responses Regarding Treatment Recommendation

Relationships Among Variables

To investigate the relationships among the variables, a series of logistic regression analyses were conducted through SPSS 6.1. In addition, an ordinal logistic regression analysis was performed using SAS 6.12 to examine hypotheses involving an ordinal dependent variable (i.e., disruptiveness). Logistic regression analysis applies a logit transformation to categorical data, so the relationship between nominal predictors (i.e., independent variables) and dichotomous outcomes (i.e., dependent variables) can be assessed. This procedure follows the same general principles as linear regression and has been noted for its ease of interpretation (Hosmer & Lemeshow, 1989).

Identification of psychological disorder

Problem type was found to be a significant predictor of identification of psychological disorder. Logistic regression analyses revealed that the log odds of identification of psychological disorder were significantly greater for cases describing aggressive disorder, b = 8.00, Wald χ2(1, N = 576) = 20.49, p < .001, for cases describing depression, b = 4.99, Wald χ2(1, N = 576) = 12.62, p < .001, and for cases describing psychosis, b = 5.31, Wald χ2(1, N = 576) = 9.47, p < .01, than for cases describing scoliosis. Furthermore, the log odds of identification of psychological disorder were significantly greater for cases describing aggressive disorder, b = 1.41, Wald χ2(1, N = 432) = 15.71, p < .001, and for cases describing psychosis, b = 1.98, Wald χ2(1, N = 432) = 21.23, p < .001, than for cases describing depression.

Client's mental retardation level, client's gender, and participant's setting were not significant predictors of identification of psychological disorder. In addition, interactions of problem type with each of these three variables did not significantly predict identification of disorder.

Recommendations for no treatment

Problem type and mental retardation level, but not gender or setting, were found to be significant predictors of a recommendation of no treatment. The log odds of this recommendation for cases describing scoliosis were significantly greater than for cases describing depression, b = −4.18, Wald χ2(1, N = 576) = 97.13, p < .001, but not significantly different from cases describing aggressive disorder or cases describing psychosis. However, these surprisingly nonsignificant Wald statistics are attributable to the large standard error for the aggressive disorder and psychosis vignettes (SE = 22.25) because no participants recommended the no-treatment option for these problem types, creating empty cells. Ninety-three percent of these recommendations were made for the scoliosis cases and 7% for cases describing depression. The log odds of this recommendation for cases describing moderate to severe mental retardation, b = 0.89, were significantly greater than for cases describing borderline to mild retardation, Wald χ2(1, N = 576) = 5.79, p < .05.

Recommendations for only on-site behavior modification

Problem type, but not mental retardation level, gender, or setting, significantly predicted a recommendation of only on-site behavior modification. The log odds of this recommendation for cases describing aggressive disorder were significantly greater than for cases describing psychosis, b = −1.23, Wald χ2(1, N = 576) = 7.23, p < .01, and for cases describing depression, b = −1.09, Wald χ2(1, N = 576) = 6.18, p < .05, but not significantly different from cases describing scoliosis.

Recommendations for verbal therapy

Problem type and mental retardation level, but not gender or setting, were significant predictors of a recommendation of verbal therapy. The log odds of this recommendation for cases describing depression were significantly greater than for cases describing aggressive disorder, b = −1.39, Wald χ2(1, N = 576) = 30.16, p < .001, cases describing psychosis, b = −2.00, Wald χ2(1, N = 576) = 53.98, p < .001, and cases describing scoliosis, b = −2.73, Wald χ2(1, N = 576) = 74.35, p < .001. The log odds of this recommendation for cases describing borderline to mild retardation were significantly greater than for cases describing moderate to severe retardation, b = −0.49, Wald χ2(1, N = 576) = 6.04, p < .05.

Recommendations for psychiatric medication

Problem type and mental retardation level, but not gender or setting, significantly predicted a recommendation of psychiatric medication. The log odds of this recommendation for cases describing depression were significantly less than for cases describing aggressive disorder, b = 2.12, Wald χ2(1, N = 576) = 14.73, p < .001, and for cases describing psychosis, b = 2.69, Wald χ2 (1, N = 576) = 24.51, p < .001, but not significantly different from cases describing scoliosis. The log odds of this recommendation for cases describing moderate to severe retardation, b = 0.85, were significantly greater than for cases describing borderline to mild retardation, Wald χ2 (1, N = 576) = 9.21, p < .01.

Recommendations for verbal therapy and medication

Problem type, but not retardation level, gender, or setting, was a significant predictor of a recommendation of verbal therapy with medication. The log odds of this recommendation for cases describing depression were significantly less than for cases describing aggressive disorder, b = 0.75, Wald χ2(1, N = 576) = 7.26, p < .01, and for cases describing psychosis, b = 1.28, Wald χ2(1, N = 576) =21.51, p < .001, and significantly greater than for cases describing scoliosis, b = −2.87, Wald χ2 (1, N = 576) = 14.94, p < .001.

Disruptiveness ratings

Problem type significantly predicted disruptiveness ratings. The log odds of the tendency toward assigning low ratings of disruptiveness for cases describing depression, b = 4.56, Wald χ2 (1, N = 432) = 150.18, p < .001, and for cases describing psychosis, b = 2.88, Wald χ2 (1, N = 432) = 89.30, p < .001, were significantly greater than for cases describing aggressive disorder.

In addition, disruptiveness ratings were found to be a significant predictor of both identification of disorder and recommendation for treatment. The log odds of identification of psychological disorder were significantly greater for cases rated as highly disruptive than for cases rated as not at all disruptive, b = −4.47, Wald χ2 (1, N = 576) = 84.05, p < .001, and for cases rated as somewhat disruptive, b = −1.00, Wald χ2 (1, N = 576) = 3.99, p < .05. Furthermore, the log odds of a recommendation for verbal therapy, medication, or both for cases rated as highly disruptive were significantly greater than for cases rated as not at all disruptive, b = −1.53, Wald χ2 (1, N = 576) = 10.50, p < .01, but not significantly different from cases rated as somewhat disruptive.

Prognosis with and without treatment ratings

Problem type was a significant predictor of ratings of prognosis with treatment. The log odds of ratings of a positive prognosis with treatment for cases describing aggressive disorder were significantly greater than for cases describing depression, b = −1.06, Wald χ2 (1, N = 432) = 4.57, p < .05, but not significantly different from cases describing psychosis.

Problem type was an even more highly significant predictor of ratings of prognosis without treatment. The log odds of a negative rating of prognosis without treatment were significantly greater for cases describing aggressive disorder than for cases describing psychosis, b = −1.42, Wald χ2(1, N = 432) = 26.46, p < .001, or for cases describing depression, b = −1.20, Wald χ2(1, N = 432) = 18.52, p < .001.

Furthermore, ratings of prognosis with and without treatment were significant predictors of a recommendation for treatment. The log odds of a recommendation of verbal therapy, medication, or both were significantly greater for cases rated as “behavior will worsen without treatment,” b = 1.26, than for cases rated as “behavior will stay the same without treatment,” Wald χ2 (1, N = 576) = 15.87, p < .001, and significantly greater for cases rated as “behavior will improve with treatment” than for cases rated as “behavior will stay the same with treatment,” b = −2.97, Wald χ2(1, N = 576) = 83.37, p < .001.

Discussion

In the current study we found that problem type and client's mental retardation level, but not worker's setting or client's gender, was related to direct-care workers' judgments. In addition, ratings of disruptiveness and of prognosis with and without treatment were related to the type of problem depicted as well as to identification of disorder and recommendations for mental health services.

Identification of Psychological Disorder and Need for Treatment

Direct-care workers were able to distinguish the presence or absence of psychological disorder in adults with mental retardation with an 89% accuracy. Furthermore, they were able to distinguish between the presence and absence of psychopathology with respect to need for treatment. A variety of treatment options were recommended, with verbal therapy and verbal therapy with medication being the most frequent. Thus, previously reported beliefs on the part of clinicians that persons with mental retardation are unable to manifest psychological disorder (Costello, 1982; Fletcher, 1988) and are not amenable to mental health intervention, particularly verbal therapy (Hurley, Pfadt, Tomasulo, & Gardner, 1996; Reiss, Levitan, & McNally, 1982), were not demonstrated by the present sample. These findings imply that although likely contributing to the underserving of the mental health needs of persons with mental retardation, faulty first-line screening for such disorders by direct-care workers is apparently not the primary determinant of this phenomenon. Direct-care workers in the present study were, however, more prone to false negative than to false positive identifications of psychological disorder. In informal assessment of psychological disorder, direct-care workers, therefore, may be more apt to underidentify than to overidentify psychopathology.

Two variables that we found to be influential in predicting direct-care workers' identification of disorder and need for treatment were problem type and mental retardation level. Results of this study supported previous suggestions that relative to externalizing and psychotic disorders, internalizing disorders may be identified less often by informants for persons with mental retardation (Foss & Peterson, 1981; Hill & Bruininks, 1984; Reid, 1977; Sovner & Hurley, 1983). Aggressive disorder and psychosis were significantly more likely than depression to be perceived as signifying the presence of psychological disorder and as necessitating treatment. Because we provided diagnostic information in the present study, recognition of internalizing disorder was independent of the client's abilities to self-report. Therefore, the finding that direct-care workers underreported depression suggests that deficient ability of persons with mental retardation to self-report symptomatology is not the sole factor responsible for the underidentification of internalizing disorder.

The present research supports the hypothesis that direct-care workers tend to selectively focus on and to seek help for behaviors that create management problems (Borthwick-Duffy & Eyman, 1990; Bruininks et al., 1988). Cases describing depression were most likely to receive low ratings of disruptiveness. Furthermore, direct-care workers were more likely to identify the presence of psychological disorder and to recommend mental health treatment for those problems that they perceived as highly disruptive than for those they perceived as mildly or not at all disruptive. Taken together, these findings suggest that direct-care workers' tendency to underreport depression as signifying psychological disorder may be partially due to the nondisruptive nature of such internalizing symptomatology. Further, cases describing depression were most likely to receive a recommendation of verbal therapy, whereas those describing psychosis or aggressive disorder were most likely to receive recommendations of psychiatric medication and verbal therapy with medication. This suggests that direct-care workers find internalizing disorders, such as depression, most amenable to verbal psychotherapy and psychosis and aggressive disorder to be most in need of treatment that includes a medication component. The increased social consequences of aggressive and psychotic symptomatology may result in an increased perception among direct-care workers that psychiatric medication will be useful or necessary in controlling these maladaptive behaviors.

Although the present research did not support the idea that a client's level of mental retardation was a significant predictor of direct-care workers' identification of psychopathology, mental retardation level was a significant predictor of perceived need for treatment. Thus, retardation level differentially impacted judgments of identification of disorder and need for treatment, such that, despite similar rates of identification of disorder, individuals with more severe mental retardation were less likely to be perceived as needing treatment. Direct-care workers may believe that individuals with more severe cognitive deficits are less able to benefit from treatment. Furthermore, this suggests that previous reports of a negative relationship between client's intellectual level and formal identification of a psychological disorder (Borthwick-Duffy & Eyman, 1990; Iverson & Fox, 1989; Jacobson, 1982, 1990b; Rojahn, 1994) may lie in differential referral and/or formal diagnostic, and not initial identification, processes.

Cases describing borderline to mild mental retardation were significantly more likely to receive a recommendation for verbal therapy, whereas those describing moderate to severe mental retardation were significantly more likely to receive a recommendation for psychiatric medication. Direct-care workers may perceive verbal therapy as less effective with individuals with more severe retardation. Furthermore, direct-care workers may perceive this subpopulation as being in greater need of medication therapy. Previous research has suggested that individuals with mental retardation often are thought to be unable to benefit from verbal psychotherapy (Hurley et al., 1996; Reiss, Levitan, & McNally, 1982). In addition, investigators have reported that the psychological problems of individuals with mental retardation frequently are believed to be of purely organic etiology and, therefore, to require psychopharmacological treatment exclusively (Cushna, Szymanski, & Tanguay, 1980; Hurley et al., 1996). The robustness of these beliefs would be expected to increase as client intellectual level decreases. The present findings support the possible presence of such beliefs among direct-care staff members, such that they are more strongly held regarding individuals with more severe cognitive deficits.

Perceptions of Prognosis With and Without Treatment

Aggressive disorder was significantly more likely than depression or psychosis to be perceived by direct-care workers as having a poor prognosis without treatment and significantly more likely than depression to be given a positive prognosis with treatment. We note that problem type was a more highly significant predictor of prognosis without treatment than of prognosis with treatment; direct-care workers appeared to strongly believe that aggressive disorder would worsen without intervention.

Further, direct-care workers were more likely to recommend verbal therapy, medication, or both for behaviors rated as worsening without treatment and for behaviors rated as improving with treatment. These findings suggest that direct-care workers' beliefs regarding the stability of maladaptive behaviors and the efficacy of mental health treatment with this population do, in fact, influence their decisions regarding treatment recommendations.

Limitations and Suggestions for Future Research

Although justified by our aim to control for client demographic characteristics across settings, the analogue nature of the current research limits the ability to generalize the findings to the judgments made by direct-care workers in actual practice. Therefore, future researchers should use natural settings to examine the impact of those variables we found to be significant predictors of direct-care workers' judgments (i.e., problem type, retardation level, disruptiveness ratings, and ratings of prognosis with and without treatment). Although sampling a greater number of employees would be desirable, in the current study this was precluded by limited personnel. The sample of 144 workers represented approximately half of direct-care workers at the participating programs. Although this sample is larger than that usually generated for such research, the extent to which the reported findings would hold for all workers is obviously uncertain. Generalizability to smaller social service agencies located in other parts of the United States is also unknown. Replication of the present methodology with differing locales and sample sizes is warranted. Future research on worker judgments also should include assessment of the contributions of occupational longevity and longevity of employment in current setting.

Research also is needed to determine whether these direct-care workers' judgments regarding treatment needs are translated into actions that increase the probability that persons with dual diagnoses will be referred for, and subsequently receive, mental health services. Consequently, in analogue research on this topic, inclusion of a measure of whether respondents would indeed initiate referrals or alert other referral sources (e.g., supervisors or family members) would add to our knowledge of the manner in which initial identification of disorder and need for treatment impact referral practices. Although empirical research has not to date established that direct-care workers' judgments impact referral practices, our clinical experience leads us to believe that such is the case and that the question merits further investigation.

Because the instruments employed here were created for the current study, further research on their reliability and validity is necessary. Although the measures appeared to generate meaningful and useful data with respect to the questions under investigation, additional support for their psychometric soundness would be desirable.

Finally, the results suggest the need to develop and evaluate staff training programs focusing on psychological disorders among persons with mental retardation. Training aimed at improving direct-care workers' ability to recognize signs of internalizing disorder would be particularly beneficial. Inclusion of information on the effectiveness, strengths, and weaknesses of various treatments appears warranted. The vignette technique of the present study could serve as one means of assessing the impact of training on workers' decision-making processes. Such training might enhance significantly the delivery of mental health services to individuals in need.

NOTE: The present study is based on the first author's doctoral research project, completed under the supervision of the second author. The authors express their appreciation to Paula Lambert for facilitating the study and to the Association for the Help of Retarded Children, the Federation of Employment and Guidance Services, and the Young Adult Institute for their participation.

References

References
Alford
,
J. D.
and
B. J.
Locke
.
1984
.
Clinical responses to psychopathology of mentally retarded persons.
American Journal of Mental Deficiency
89
:
195
197
.
Benson
,
B. A.
1985
.
Behavior disorders and mental retardation: Associations with age, gender, and level of functioning in an outpatient clinic sample.
Applied Research in Mental Retardation
6
:
79
85
.
Borthwick
,
S. A.
1988
.
Maladaptive behavior among the mentally retarded: The need for reliable data.
In J. A. Stark, F. J. Menolascino, M. H. Albarelli, & V. C. Gray (Eds.), Mental retardation and mental health: Classification, diagnosis, treatment, services (pp. 30–40). New York: Springer-Verlag
.
Borthwick-Duffy
,
S. A.
1994
.
Epidemiology and prevalence of psychopathology in people with mental retardation.
Journal of Consulting and Clinical Psychology
62
:
17
27
.
Borthwick-Duffy
,
S. A.
and
R. K.
Eyman
.
1990
.
Who are the dually diagnosed?
American Journal on Mental Retardation
94
:
586
595
.
Bruininks
,
R. H.
,
B. K.
Hill
, and
L. E.
Morreau
.
1988
.
Prevalence and implications of maladaptive behaviors and dual diagnosis in residential and other service programs.
In J. A. Stark, F. J. Menolascino, M. H. Albarelli, & V. C. Gray (Eds.), Mental retardation and mental health: Classification, diagnosis, treatment, services (pp. 3–29). New York: Springer-Verlag
.
Costello
,
A.
1982
.
Assessment and diagnosis of psychopathology.
In J. L. Matson & R. Barrett (Eds.), Psychopathology in the mentally retarded (pp. 37–52). New York: Grune & Stratton
.
Cushna
,
B.
,
L. S.
Szymanski
, and
P. E.
Tanguay
.
1980
.
Professional roles and unmet manpower needs.
In L. S. Szymanski & P. E. Tanguay (Eds.), Emotional disorders of mentally retarded persons: Assessment, treatment, and consultation (pp. 3–18). Baltimore: University Park Press
.
Eaton
,
L. F.
and
F. J.
Menolascino
.
1982
.
Psychiatric disorders in the mentally retarded: Types, problems, and challenges.
American Journal of Psychiatry
139
:
1297
1303
.
Edelstein
,
T. M.
and
D. S.
Glenwick
.
1997
.
Referral reasons for psychological services for adults with mental retardation.
Research in Developmental Disabilities
18
:
45
59
.
Fletcher
,
R. J.
1988
.
County systems model: Comprehensive services for the dually diagnosed.
In J. A. Stark, F. J. Menolascino, M. H. Albarelli, & V. C. Gray (Eds.), Mental retardation and mental health: Classification, diagnosis, treatment, services (pp. 254–264). New York: Springer-Verlag
.
Fletcher
,
R. J.
1993
.
Mental illness-mental retardation in the United States: Policy and treatment challenges.
Journal of Intellectual Disability Research
37
:
25
33
.
Foss
,
G.
and
S.
Peterson
.
1981
.
Social-interpersonal skills relevant to job tenure for mentally retarded adults.
Mental Retardation
19
:
103
106
.
Hill
,
B.
and
R.
Bruininks
.
1984
.
Maladaptive behavior of mentally retarded people in residential facilities.
American Journal of Mental Deficiency
88
:
380
387
.
Hosmer
,
D. W.
and
S.
Lemeshow
.
1989
.
Applied logistic regression.
New York: Wiley
.
Hurley
,
A. D.
,
A.
Pfadt
,
D.
Tomasulo
, and
W. I.
Gardner
.
1996
.
Counseling and psychotherapy.
In J. W. Jacobson & J. A. Mulick (Eds.), Manual of diagnosis and professional practice in mental retardation (pp. 371–380). Washington, DC: American Psychological Association
.
Iverson
,
J. C.
and
R. A.
Fox
.
1989
.
Prevalence of psychopathology among mentally retarded adults.
Research in Developmental Disabilities
10
:
77
83
.
Jacobson
,
J. W.
1982
.
Problem behavior and psychiatric impairment within a developmentally disabled population: 1. Behavior frequency.
Applied Research in Mental Retardation
3
:
121
139
.
Jacobson
,
J. W.
1990a
.
Assessing the prevalence of psychiatric disorders in a developmentally disabled population.
In E. Dibble & D. B. Gray (Eds.), Assessment of behavior problems in persons with mental retardation living in the community (pp. 19–70). Rockville, MD: U. S. Department of Health and Human Services
.
Jacobson
,
J. W.
1990b
.
Do some mental health disorders occur less frequently among persons with mental retardation?
American Journal on Mental Retardation
94
:
596
602
.
Lopez
,
S. R.
1989
.
Patient variable biases in clinical judgment: Conceptual overview and methodological considerations.
Psychological Bulletin
106
:
184
203
.
Menolascino
,
F. J.
,
S. F.
Gilson
, and
A. S.
Levitas
.
1986
.
Issues in the treatment of mentally retarded patients in the community mental health system.
Community Mental Health Journal
22
:
314
327
.
Nezu
,
C. M.
and
A. M.
Nezu
.
1994
.
Outpatient psychotherapy for adults with mental retardation and concomitant psychopathology: Research and clinical imperatives.
Journal of Consulting and Clinical Psychology
62
:
34
42
.
Pawlarcyzk
,
D.
and
B. E.
Beckwith
.
1987
.
Depressive symptoms displayed by persons with mental retardation: A review.
Mental Retardation
25
:
325
330
.
Reid
,
A. H.
1977
.
Psychosis in adult mental defectives: I. Manic depressive psychosis.
British Journal of Psychiatry
120
:
205
212
.
Reiss
,
S.
1982
.
Psychopathology and mental retardation: Survey of a developmental disabilities mental health program.
Mental Retardation
20
:
128
132
.
Reiss
,
S.
1990
.
Prevalence of dual diagnosis in community-based programs in the Chicago metropolitan area.
American Journal on Mental Retardation
94
:
578
585
.
Reiss
,
S.
1994
.
Reiss Screen for Maladaptive Behavior: Test manual (2nd ed.).
Worthington, OH: IDS.
Reiss
,
S.
,
G. W.
Levitan
, and
R. J.
McNally
.
1982
.
Emotionally disturbed mentally retarded people: An underserved population.
American Psychologist
37
:
361
367
.
Reiss
,
S.
,
G. W.
Levitan
, and
J.
Szyszko
.
1982
.
Emotional disturbance and mental retardation: Diagnostic overshadowing.
American Journal of Mental Deficiency
86
:
567
574
.
Reiss
,
S.
and
J.
Szyszko
.
1983
.
Diagnostic overshadowing and professional experience with mentally retarded persons.
American Journal of Mental Deficiency
87
:
396
402
.
Rojahn
,
J.
1994
.
Epidemiology and topographic taxonomy of self-injurious behavior.
In T. Thompson (Ed.), Destructive behavior in developmental disabilities: Diagnosis, measurement, and evaluating treatment outcome (pp. 49–67). Newbury Park, CA: Sage
.
Rojahn
,
J.
and
M. J.
Tassé
.
1996
.
Psychopathology in mental retardation.
In J. W. Jacobson & J. A. Mulick (Eds.), Manual of diagnosis and professional practice in mental retardation (pp. 147–156). Washington, DC: American Psychological Association
.
Rutter
,
M.
and
P.
Graham
.
1970
.
Epidemiology of psychiatric disorder.
In M. Rutter, J. Tizard, & K. Whitmore (Eds.), Education, health, and behavior (pp. 178–201). London, England: Longman Group
.
SAS
6.12 [Computer software]
.
1996
.
SAS Institute.
Cary, NC: Author
.
Sovner
,
R.
and
A.
Hurley
.
1983
.
Do the mentally retarded suffer from affective illness?
Archives of General Psychiatry
40
:
61
67
.
Spengler
,
P. M.
,
D. C.
Strohmer
, and
H. T.
Prout
.
1990
.
Testing the robustness of the diagnostic overshadowing bias.
American Journal on Mental Retardation
95
:
204
214
.
SPSS
6.1 [Computer software]
.
1994
.
SPSS.
Chicago: Author
.
Szymanski
,
L. S.
1980
.
Individual psychotherapy with retarded persons.
In L. S. Szymanski & P. E. Tanguay (Eds.), Emotional disorders of mentally retarded persons: Assessment, treatment, and consultation (pp. 131–148). Baltimore: University Park Press
.

Author notes

Authors:Teri M. Edelstein, PhD, Clinical Psychologist, Department of Psychiatry, New York University/Bellevue Hospital Center, 27th St. and First Ave., New York, NY 10016. David S. Glenwick, PhD, Professor, Department of Psychology, Fordham University, 441 E. Fordham Rd., Bronx, NY 10458–9993. Requests for reprints should be sent to the second author.