Full Text of: CROGHAN: J Nerv Ment Dis, Volume 191(3).March 2003.166-174
The Journal of Nervous and Mental Disease (C) 2003 Lippincott Williams & Wilkins, Inc.
----------------------------------------------Volume 191(3) March 2003 pp 166-174 ----------------------------------------------
AMERICAN ATTITUDES TOWARD AND WILLINGNESS TO USE PSYCHIATRIC MEDICATIONS [Articles] CROGHAN, THOMAS W. M.D.1; TOMLIN, MOLLY M.S.2; PESCOSOLIDO, BERNICE A. Ph.D.3; SCHNITTKER, JASON Ph.D.4; MARTIN, JACK Ph.D.5; LUBELL, KERI Ph.D.3; SWINDLE, RALPH Ph.D.2 1 Department of Sociology, Indiana University, Bloomington, Indiana and RAND Corporation, Arlington, VA. Send reprint requests to Dr. Croghan, Department of Sociology, c/o Indiana Consortium for Mental Health Services Research, Indiana University, 1022 East Third St., Bloomington, IN 47405. 2 Eli Lilly and Company, Indianapolis, Indiana. 3 Department of Sociology, Indiana University, Bloomington, Indiana. 4 Department of Sociology, University of Pennsylvania, Philadelphia, Pennsylvania. 5 Department of Sociology, Kent State University, Kent, Ohio. Supported by grants from the National Science Foundation, the National Institute of Mental Health, the National Institute on Alcohol Abuse and Alcoholism, and Eli Lilly and Company.
Study Population Dependent Variable Measures
Willingness to Use Prescribed Psychiatric Medications.
Independent Variable Measures
Attitudes Toward Psychiatric Medications. Sociodemographic Attributes and Health Status. Attitudes Toward Physicians. Analyses.
Discussion Conclusion References
Table 1 Table 2 Table 3 Table 4
Despite recent advances in treatment, many Americans decline to take prescribed psychiatric medication. This study explores the role of attitudes regarding the effectiveness of and potential problems associated with psychiatric medications on Americans' willingness to use them. Face-to-face interviews of a US household population sample were done with 1387 volunteers. The 1998 General Social Survey's Pressing Issues in Health and Medical Care Module (response rate, 76.4%) included questions about efficacy, problems, and potential use. Most Americans agree that psychiatric medications are effective, and fewer than half had concerns regarding potential problems. However, the majority of respondents would not be willing to take them. Willingness to use is influenced by these attitudes and other factors, including health status and past use of mental health treatments. Although Americans perceive psychiatric medications to be effective, and this influences their willingness to take them, many still are not willing to take them.
Approximately 44 million Americans are believed to have a diagnosable mental disorder in any year (Hirschfeld et al., 1997;Kessler et al., 1994;Regier et al., 1993), and there are many effective treatments for these problems (Mental Health: A Report of the Surgeon General, 1999). In particular, newer psychiatric medications have made treatment more tolerable and easier to deliver than in the past. These new medications, especially the selective serotonin reuptake inhibitors (SSRIs) and the atypical antipsychotic agents, have been widely accepted by the medical community and are frequently prescribed, fueling public debate regarding the benefits, risks, and costs of mental health treatments.
Despite established need and treatment efficacy, studies document that less than one third of persons believed to need mental health treatment actually receive it (Hirschfeld et al., 1997). Most theoretical models of health service utilization, including the Health Belief Model, the Theory of Reasoned Action, and the Socio-Behavioral Model, conceptualize attitudes about perceived efficacy as central to decisions that persons make about going for care and following through on physician-recommended treatments (Ajzen and Fishbein, 1980;Andersen, 1995;Rosenstock, 1966). Consistent with these models, the Surgeon General's Report points to attitudes and lack of understanding regarding the range of effective treatment choices as barriers to mental health care (Mental Health: A Report of the Surgeon General, 1999).
To our knowledge, the single body of past US research on the awareness of and reactions to psychiatric medication assessed public responses to tranquilizers during a period in which public controversy about their use was high, i.e., the 1970s (Clinthorne et al., 1986;Manheimer et al., 1973;Mellinger et al., 1978). This research suggested that public understanding of effectiveness stood in contrast to an unwillingness to take these drugs, reflecting concerns about long-term effects and the "morality" of pharmacological solutions to behavioral problems. These older US studies contrast with more recent European studies in which psychiatric medications were generally perceived as dangerous, unnatural, and harmful (Angermeyer et al., 1993;Horne, 1999;Slovick et al., 1989). For example, in Sweden, antidepressants were associated with higher perceptions of risk than heart surgery (Slovick et al., 1989). These beliefs influenced willingness to use, and subsequent adherence to, medication regimens (Horne, 1999).
Since the original US studies, there has been substantial debate about community attitudes and acceptance of psychiatric medications, with some arguing for a growing sophistication among the public and others decrying the widespread use of psychiatric medications, particularly the SSRIs (Glenmullen, 2000;Mental Health: A Report of the Surgeon General, 1999). These discussions about public attitudes and the low level of mental health service use have occurred in the absence of data on whether the gap between available treatments and mental health service use results from ignorance, lack of acceptance of scientific evidence, or other factors.
Recent research demonstrates that Americans correctly identify mental disorders such as schizophrenia and depression (Link et al., 1999). Furthermore, Americans report that mental disorders are caused by multiple factors, including genetic factors, stress, and biological phenomena, consistent with current professional and scientific understanding. They also indicate that situations describing diagnosable mental health disorders will not improve on their own but will do so with treatment, highly endorsing the use of a wide range of general and specialty providers (Pescosolido et al., 2000).
Despite apparent knowledge and understanding, the same Americans who reported these positive beliefs also reported low levels of potential service use if they anticipated a mental health problem in their own lives (Swindle et al., 2000). This recent series of studies did not ask, in detail, about Americans' attitudes toward and acceptance of one of the primary treatment options in contemporary psychiatry-medications. In 1998, the General Social Survey (GSS), the nation's leading monitor of social and political attitudes, followed up on the earlier studies of attitudes toward mental disorders with a specific set of questions about the generic category of psychiatric medications. These questions tapped public attitudes regarding effectiveness, potential problems, and the likelihood of personally using these medications when prescribed under specific circumstances.
The purpose of this study is to use the 1998 GSS data to address four important questions for mental health care and mental health policy. First, how do Americans assess the effectiveness of and problems with psychiatric medications? Second, how likely and under what conditions do Americans report a willingness to use medications? Third, how is willingness influenced by perceived efficacy? Fourth, how is this relationship modified by concerns about psychiatric medications, sociodemographic and socioeconomic characteristics, personal health-related factors and experiences, and evaluations of medical system providers? These questions set the larger context in which the public responds to mental health problems, weighs individual costs and benefits of seeking care, and creates the social influence pushing people to or away from formal treatment (Pescosolido, 1991;Pescosolido and Boyer, 1999).
Data come from the Pressing Issues in Health and Medical Care Module (PIHMC) of the 1998 GSS conducted by the National Opinion Research Center, University of Chicago (Davis et al., 1998). The GSS is a biennial, 90-minute face-to-face interview of US households. It includes three types of questions: basic sociodemographic questions on every survey, rotating "core" questions asked of two of three respondents in each fielding, and one-time questions asked on "topical modules." The 1998 GSS used a 3-stage full probability sampling design, reporting data from two representative samples of the English-speaking, adult, noninstitutionalized population of the contiguous United States (N = 2832). Analyses reported here are based on the responses of 1387 persons in one sample who were administered the 73-item PIHMC module. Data were collected in February and March 1998 (overall response rate, 76.4%). The 1998 GSS received IRB approval from the University of Chicago.
Dependent Variable Measures
Willingness to Use Prescribed Psychiatric Medications.
PIHMC respondents were asked about four conditions: a) having troubles in one's personal life; b) the ability to cope with stresses of life; c) feeling depressed, tired, having trouble sleeping and concentrating, and feelings of worthlessness; and d) having periods of intense fear associated with trembling, sweating, feeling dizzy, and fear of losing control or going crazy. Respondents were asked, for each, whether they were very likely, somewhat likely, unsure, somewhat unlikely, or very unlikely to take physician-prescribed psychiatric medication. For descriptive analyses, we collapsed responses into "likely," "unlikely," and "unsure." Respondents were not offered a "don't know" or "no answer" option, but these responses were coded as such if they were offered spontaneously.
To construct a multi-item measure, willingness to use, for the multivariate analyses, unidimensionality of the willingness items was assessed via principal components factor analysis. Examination of intercorrelations among the nonrecoded items indicated a single underlying dimension. All items produced factor loadings greater than .80 with no evidence of double loading (details on request). We combined the four items into a composite indicator of the willingness to use prescribed psychiatric medications. This composite weights each component's contribution via that item's factor score coefficient. The final measure reports the 4-component sum after z-score transformations with an internal consistency reliability of .88 (Cronbach's alpha). Higher scores indicate greater willingness to use.
Independent Variable Measures
Attitudes Toward Psychiatric Medications.
This was measured by six items that assessed respondents' reports of perceived effectiveness (i.e., psychiatric medications help with day-to-day stress, make relations with family and friends easier, help control symptoms, and help people feel better about themselves) and concerns regarding potential problems (i.e., psychiatric medications are harmful to the body and interfere with daily activities;Streiker et al., 1986). Respondents were asked to strongly agree, agree, neither agree nor disagree, disagree, or strongly disagree with each item. For the descriptive analyses, we collapsed responses into "agree," "disagree," and "neither agree nor disagree." The "don't know" and "no answer" options were again not presented but were coded as such if offered spontaneously.
Principal components factor analyses of the original 7-item battery assessing attitudes toward psychiatric medications indicated that items tapped the two anticipated conceptual dimensions. Four effectiveness items loaded on one factor (all item loadings greater than .65, no evidence of "double loading"), while two items captured concerns about side effects (loadings greater than .76, no evidence of "double loading"). A seventh item, belief that taking psychiatric medications should stop after symptoms were no longer present, did not load on either factor and was eliminated from further consideration. Thus, for model fitting, we combined the four effectiveness items and two concern items into composite measures with an item's contribution to the scale weighted via the use of factor score coefficients. Each resultant scale represents the sum, after z-score transformation, of the component items. Internal consistency reliabilities for the two scales are .80 and .61, respectively. Higher scale scores indicate greater levels of that construct.
Sociodemographic Attributes and Health Status.
Sociodemographic attributes utilized include age measured in years and log income taken as the natural log of self-reported 1998 earnings. Education was measured with a series of binary variables coded 0 or 1, respectively, if the respondent had graduated from high school or had some college education (high school-some college), completed a college degree, or had a graduate degree. Respondents with less than high school education served as the omitted category. Gender and race were also coded as dichotomies (women, 0 = male, 1 = female; and whites, 1 = white, 0 = other), and place of residence and marital status were coded as dummy variable sets. Residence was coded into rural and urban categories with suburban residents as the reference group. Marital status was composed of three binary variables-widowed, divorced-separated, and never married. Married subjects constituted the omitted category.
The PIHMC module includes health status-related questions. Self-reported health status was assessed by a single question, a 4-point ordinal metric ranging from poor, fair, good, to excellent. Health insurance status, insured, was assessed with a single item where respondents were coded 1 if they reported holding public or private health insurance or 0, if otherwise. Subjects were coded 1 if they indicated that their health plan required them to select doctors from a list (0, otherwise). Finally, respondents were coded as 1 if they reported that they had used mental health services or knew someone who had (self-others used MHS).
Attitudes Toward Physicians.
Attitudes regarding physicians and the health care system were assessed by 20 items drawn from the work of Aday et al. (1980), Kao et al. (1998), Anderson and Dedrick (1990), and Thom et al. (1999). Initial factor analyses indicated that these items tap four distinct dimensions: a) trust in one's personal physician (7 items); b) worries about the denial of needed care (3 items); c) overall confidence in physicians (5 items); and d) general concern about physician quality (4 items). Consistent with scaling procedures described previously, composite indices of each of these conceptual dimensions were constructed using factor score coefficients to weight component items, with resultant scale values representing the sums, after z-score transformation, of the component items. They ranged from low to high with internal consistency reliabilities, assessed by Cronbach's alpha, ranging from .67 to .83.
Initial analyses to recode variables and establish multi-item scales were conducted with the 1387 unweighted subjects. In subsequent multivariate analyses, to correct for design effects associated with unequal probabilities of selection related to the clustering of GSS sample elements within primary sampling units (PSUs), we utilized SUDAAN to weight the data. All model estimations reported on the weighted data and utilized procedures to adjust coefficients and standard errors to take the complex survey design into account (GSS sample weights provided on request by GSS/NORC sampling staff;Binder, 1983). Finally, to avoid erroneous inferences derived from dropping respondents with missing data on multi-item scales when missing responses to component items were not random occurrences, we substituted the mean variable score when a subject was missing data on no more than one scale item for all scales except the "trust one's personal physician" scale, when we substituted the mean value when three or fewer responses were missing. This substitution was used for 13 cases on the willingness to use scale, 32 cases on the perceived effectiveness scale, 46 cases on the trust one's personal physician scale, 13 cases on the worries about denial scale, 32 cases on the overall confidence in one's physician scale, and 41 cases on the general concerns about physician quality scale. We also used a regression-based imputation procedure for 82 respondents who did not report income.
The specific research questions posed here were addressed with a 2-step analysis. First, we evaluated whether the public is willing to use prescribed psychiatric medications under various circumstances, whether they perceive these medications to be effective, and whether they believe use of these medications is associated with potential problems. These descriptive analyses used univariate percentage distributions. Second, to examine the relationship between the willingness to use psychiatric medications and various independent variables, we estimated a series of weighted least squares regression models where explanatory variables are entered in sequenced sets. To begin, we regressed willingness to use psychiatric medications on the perceived effectiveness scale, examining the zero-order relationship. Next, the potential problems scale was entered in a second step, and that model was then supplemented by estimating the effects of sociodemographic characteristics (gender, race, education, income, urban/rural residence, marital status, and age). Next, several health status-related variables focusing on personal circumstances (e.g., self-reported health status, insurance coverage) were included. Finally, to examine whether the relationship between perceived effectiveness and willingness to use is influenced by more general sentiments about providers in the health care system, we added attitudes toward physicians (general and personal) and system access (worries about the denial of treatment) to the model. We report unstandardized regression coefficients and one-tailed tests of statistical significance.
Table 1 indicates that the profile of GSS respondents generally mirrors the sociodemographic characteristics of the US adult population. With the exception of gender, where there appears to be a slight overrepresentation of women (58.1%), the sociodemographic profile in this GSS sample is well within ranges representative of the US population. Also consistent with national trends, 86% of PIHMC respondents report having some form of health insurance, and of those with health insurance, approximately 60% must select their health care provider from a list. The majority of respondents rate their general health status as good (48.6%) or excellent (30.6%), with fewer than one in four rating their health as fair or poor. Finally, more than half (59%) report that either they personally have seen or know someone who has seen a psychiatrist, psychologist, or counselor.
----------------------------------------------TABLE 1 Characteristics of the PIHMC sample (n = 1,369)a Asked only of those with health insurance. ----------------------------------------------
Table 2 reports Americans' willingness to take psychiatric medications for four mental health problems. Except for the situation involving intense physical and emotional symptoms ("going crazy"), more respondents said they would be unlikely to take prescribed psychiatric medication. Only 24% said they were likely to take psychiatric medication in response to trouble in personal life, while 64.4% were unlikely to do so. More than one third (36.7%) were likely to take psychiatric medication to cope with stress, but almost half (49.9%) would not. For depression, Americans were about evenly split with 41.3% reporting they would likely take psychiatric medications, while 45.4% would not. Finally, more than half of respondents (56.0%) would take psychiatric medications if they believed that they were "going crazy" and having periods of intense fear and physical symptoms. However, even under these circumstances, 30.3% of Americans indicated they would not take prescribed psychiatric medications.
----------------------------------------------TABLE 2 Item distributions (%) on likelihood of taking psychiatric medications, 1998 general social survey, PIHMC module ----------------------------------------------
More than 1 in 10 PIHMC respondents reported being unsure whether they would take psychiatric medication for each of the situations presented. Further, cumulative results not shown here indicate that more than one third of respondents (37.1%) report being unlikely to take psychiatric medications under any of the circumstances. Only 16.8% reported being likely to take psychiatric medications in one of the situations, 14.1% in two, 15.1% in three, and only 18.9% in all four.
Table 3 presents Americans' attitudes about the efficacy of and concerns regarding potential problems. Overwhelmingly, Americans see psychiatric medication as effective. More than three of four respondents agreed that psychiatric medications help control symptoms (77.2%) and help people deal with daily stresses (72.6%). Somewhat fewer (61.7%) believed that psychiatric medications made relationships easier, and about half (54.9%) agreed that psychiatric medications made people feel better about themselves. In cumulated results not shown here, only 9% of respondents failed to agree with at least one of the statements regarding the effectiveness of psychiatric medications, while 42% agreed with all four statements.
----------------------------------------------TABLE 3 Item distributions (%) on attitudes toward psychiatric medications, 1998 general social survey, PIHMC modulea Strongly agree, agree combined; strongly disagree, disagree combined. ----------------------------------------------
Americans simultaneously expressed concern about potential problems associated with use of psychiatric medications, with 23.3% of respondents agreeing that psychiatric medications were harmful to the body and 35.8% agreeing that these medications interfered with daily activities (Table 3). From 10% to 20% of subjects responded they were uncertain about problems associated with use of psychiatric medications, depending on the specific question. Few respondents (from 7% to 13%) reported that they did not know or provided no answer. Further, in cumulated results not shown here, more than half (51%) did not endorse either of the potential concerns about psychiatric medications. Almost one third (30.2%) agreed with one of the statements of concern, and 18.8% agreed with both statements.
Tables 2 and 3, taken together, indicate that Americans tend to be unwilling to take physician-prescribed psychiatric medications even as they maintain generally positive attitudes toward them. To explore this gap, we analyzed the relationship between the two. Further, we examined four sets of factors that might mediate or modify the relationship between perceived effectiveness and willingness to use these medications. As described above, these include concerns regarding potential problems associated with use of psychiatric medications, sociodemographic characteristics, health and health insurance status, and general attitudes toward physicians and medical care.
Table 4 indicates, as expected by most theoretical utilization models, attitudes toward psychiatric medications are strongly associated with likely use. Respondents with positive attitudes regarding the effectiveness of psychiatric medication were significantly more likely to report willingness to use them ([beta] = .395, p
----------------------------------------------TABLE 4 Weighted least squares estimates for the regression of overall likelihood of using psychiatric medications on attitudes toward psychiatric medications, background attributes, health status variables, and attitudes toward physicians, 1998 general social survey, PIHMC module*** p ----------------------------------------------
Column 2 adds an estimate for the influence of concerns regarding potential problems to the model. As anticipated, the impact of these beliefs was negative and statistically significant ([beta] = -.192, p i.e., 3.3%) and did little to alter the magnitude of the observed positive impact of perceived effectiveness.
In an attempt to examine evidence of any social, cultural, or economic differences in respondents' willingness to use prescribed psychiatric medications, column 3 estimates the impact of sociodemographic attributes. Educational attainment had a powerfully negative effect on willingness to use psychiatric medications. Compared with their counterparts with fewer than 12 years of schooling, better-educated respondents, particularly those with 4-year college degrees ([beta] = -.321, p p p
Column 4 assesses the effects of four health system or health status-related dimensions. Respondents with familiarity with the mental health system, either through personal use or knowledge of others' consultation, were significantly more likely to report a willingness to use psychiatric medications ([beta] =.181, p p
Column 5 assesses the extent to which the relationship between perceived effectiveness and willingness to use is influenced by general sentiments about physicians (general and personal, including trust) and access to the health care system (worries about the denial of treatment). Respondents who reported more trust in personal physicians ([beta] = .082, p p
The data presented here suggest that Americans have generally positive attitudes toward psychiatric medications, with the majority agreeing with statements regarding effectiveness. There also appears to be substantial concern regarding potential problems with their use since one quarter of respondents believe that psychiatric medications are harmful to the body. Overall, these findings suggest an acceptance of psychiatric medication tinged with a substantial skepticism.
Like US studies from the 1970s, we find that subjects report that they believe that psychiatric medications are effective at the same time that they indicate that they are unlikely to use them. This contrasts with European-based studies in which perceptions of effectiveness and safety appear to more closely track willingness to use the medication. These studies report more negative attitudes and a more consistent low willingness to use psychiatric medication.
The US juxtaposition of generally positive attitudes regarding effectiveness and low levels of willingness to use prescribed psychiatric medication under circumstances that describe common mental disorders suggests a critical gap. If Americans believe or hold the opinion that psychiatric medications are effective, why are they unwilling to use them? In the contemporary United States, we find a significant but only moderate link between attitudes and willingness to use. Concerns with these medications explain only some of this unwillingness, and sociodemographic characteristics explain little of the variation in either perceived effectiveness (in results not reported here) or willingness to use. Several factors traditionally considered barriers to care appear to have little effect on willingness to use psychiatric medications. Specifically, gender, race, rural-urban residence, having health insurance, overall assessment of physicians, and being forced to choose doctors from a list were not associated with willingness to use prescribed psychiatric medication. In other words, the gap described here is not explained by differences based on gender, race, or urban living.
Unfortunately, the research presented here does not explain the gap between attitudes and willingness to use prescribed psychiatric medications, but there may be other explanations that cannot be addressed with the available data. First, recent research suggests that Americans are more likely to seek mental health care from counselors and other nonphysicians than they were in the past (Swindle et al., 2000). Thus, low willingness to use medications may reflect positive preferences for certain types of psychosocial treatments and providers, particularly for those with higher levels of education. Second, the gap may be explained in part by the stigma associated with mental illness that evokes fear and embarrassment about disclosure, deterring persons from seeking treatment (Mental Health: A Report of the Surgeon General, 1999).
There is a third, more positive, interpretation of the results presented here. Our findings may reflect attitudes about autonomy, empowerment, and control that have been found to be important to the outcomes of those with chronic diseases such as diabetes (Bandura and Simon, 1981). Consistent with this theory, many Americans report preferences for nonmedical treatment or self-treatment when faced with symptoms commonly associated with mental disorders (Pescosolido et al., 2000), and most experts agree that dealing with any underlying cause of psychological distress is important to long-term outcome. In this interpretation, American self-image and its corresponding potential for self-efficacy may contrast with the rhetoric of experts who suggest wider use of psychiatric medications. Further research is necessary to assess the effects of these factors on willingness to use psychiatric medications.
Our study has important implications for health care policy and the current debate regarding increases in spending for pharmaceuticals in general and psychiatric medications in particular. Many analysts suggest that factors such as rich pharmacy benefits and direct to consumer advertising have encouraged unnecessary use of the medications (Croghan, 2001). Although unnecessary use may coexist with low levels of treatment for those in need, our research adds to the circumstantial evidence that factors such as insurance and advertising may not result in unnecessary use of psychiatric medications. Indeed, the positive image of tranquilizers and antidepressants presented by television advertising appears to have been known and understood prior to broad use of this advertising. It remains to be determined whether these advertisements have altered willingness to use psychiatric medications.
Inferences based on analyses presented here are limited by several factors. The 1998 GSS overrepresents women, potentially introducing gender bias into the descriptive results. Women are more likely than men to recognize a mental health problem, be diagnosed with a mental disorder (Kessler et al., 1994), and seek medical care for either a physical or mental health problem (Leaf et al., 1988). These factors may suggest that, even in the face of the gap between attitudes and predisposition, we may overestimate willingness to use psychiatric medications. Further, we do not know the respondents' referent for the general category of psychiatric medications described in the conditions. Although one may hypothesize that recent public discussion regarding use of antidepressant medication would play prominently in the minds of PIMHC respondents, others with greater knowledge or contact with persons with serious mental disorders such as schizophrenia may have other medications in mind. Others hypothesize that the public generalizes the term "psychiatric medications" to mean the benzodiazepines, especially focusing on their side effect profile (Jorm, 2000).
The data presented here replicate a gap seen in earlier studies-Americans believe that psychiatric medications are effective at relieving symptoms associated with mental disorders, yet they are relatively unwilling to use them in most situations. These findings are consistent with research documenting the discrepancy between the level of need for mental health services and the actual use of medical treatments. While our results support the theoretical link between attitudes and willingness, the gap between attitudes and willingness to use prescribed psychiatric medications remains to be fully explained.
1. Aday L, Andersen R, Fleming G ( 1980) Health care in the United States. Beverly Hills: Sage.
2. Ajzen I, Fishbein M ( 1980) Understanding attitudes and predicting behavior. Englewood Cliffs, NJ: Prentice Hall.
3. Andersen R ( 1995) Revisiting the behavioral model and access to care: Does it matter? J Health Soc Behav 36: 1-10. [Medline Link]
4. Anderson R, Dedrick R ( 1990) Development of the trust in physician-scale: A measure to assess interpersonal trust in patient-physician relationships. Psychol Rep 76: 1091-1100. [Medline Link] [BIOSIS Previews Link]
5. Angermeyer MC, Daumer R, Matschinger H ( 1993) Benefits and risks of psychotropic medication in the eyes of the general public: Results of a survey in the Federal Republic of Germany. Pharmacopsychiatry 26: 114-120. [Medline Link] [BIOSIS Previews Link]
6. Bandura A, Simon K ( 1981) Cultivating competence, self-efficacy, and intrinsic interest through proximal self-motivation. J Pers Soc Psychol 41: 586-598.
7. Binder DA ( 1983) On the variances of asymptotically normal estimators from complex data. Int Stat Rev 51: 279-292.
8. Clinthorne JK, Cisn IH, Balter MB, Mellinger GD, Uhlenhuth EH ( 1986) Changes in popular attitudes and beliefs about tranquilizers. Arch Gen Psychiatry 43: 527-532. [Medline Link] [BIOSIS Previews Link]
9. Croghan TW ( 2001) The controversy of rising antidepressant expenditures. Health Aff (Millwood) 20: 129-135. [Medline Link]
10. Davis JA, Smith TW, Marsden PV ( 1998) General social surveys: 1972-1998. Chicago: National Opinion Research Center.
11. Glenmullen J ( 2000) Prozac backlash. New York: Simon and Schuster.
12. Hirschfeld RM, Keller MB, Panico S, Arons BS, Barlow D, Davidoff F, Endicott J, Fromm J, Goldstein M, Gorman JM, Marek RG, Maurer TA, Meyer R, Phillips K, Ross J, Schwenk TL, Sharfstein SS, Thase ME, Wyatt RJ ( 1997) National Depressive and Manic Depressive Association consensus statement on the undertreatment of depression. JAMA 277: 333-340. [Fulltext Link] [Medline Link] [CINAHL Link] [BIOSIS Previews Link]
13. Horne R ( 1999) Patients' beliefs about treatment: The hidden determinant of treatment outcome? J Psychosom Res 6: 491-495. [Medline Link]
14. Jorm AF ( 2000) Mental health literacy: Public knowledge and beliefs about mental disorders. Br J Psychiatry 177: 396-401. [Medline Link] [BIOSIS Previews Link]
15. Kao AC, Green DC, Davis NA, Koplan JP, Cleary PD ( 1998) Patients' trust in their physicians: Effects of choice, continuity, and payment method. J Gen Intern Med 13: 681-686. [Medline Link]
16. Kessler RC, McGonagle KA, Zhao S, Nelson CD, Hughes M, Eshleman S, Wittchen H-U, Kendler KS ( 1994) Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Results from the National Comorbidity Survey. Arch Gen Psychiatry 51: 8-19. [Fulltext Link] [Medline Link] [BIOSIS Previews Link]
17. Leaf PJ, Bruce ML, Tischler GL, Freeman DH, Weismann MM, Myers JK ( 1988) Factors affecting the utilization of specialty and general medical mental health services. Med Care 26: 9-26. [Medline Link]
18. Link BG, Phelan JC, Bresnahan M, Stueve A, Pescosolido BA ( 1999) Public conceptions of mental illness: Labels, causes, dangerousness, and social distance. Am J Public Health 89: 1328-1333. [Medline Link] [CINAHL Link] [BIOSIS Previews Link]
19. Manheimer DI, Davidson ST, Balter MB, Mellinger GD, Cisin IH, Parry HJ ( 1973) Popular attitudes and beliefs about tranquilizers. Am J Psychiatry 130: 1246-1253. [Medline Link]
20. Mellinger GD, Balter MB, Manheimer DI, Cisin IH, Parry HJ, ( 1978) Psychic distress, life crisis, and use of psychotherapeutic medications. Arch Gen Psychiatry 35: 1045-1052. [Medline Link]
21. Mental Health: A Report of the Surgeon General ( 1999) Rockville, MD: US Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, Center for Mental Health Services, National Institutes of Health, National Institute of Mental Health.
22. Pescosolido BA ( 1991) Illness careers and network ties: A conceptual model of utilization and compliance. Adv Med Sociol 2: 161-184.
23. Pescosolido BA, Boyer CA ( 1999) How do people come to use mental health services? Current knowledge and changing perspectives. In Horwitz AV, Scheid TL (Eds), A handbook for the study of mental health. New York: Cambridge University Press. 441-460.
24. Pescosolido BA, Martin JK, Link BG, Kikuzawa S, Burgos G, Swindle R, Phelan J ( 2000) Americans' views of mental health and illness at century's end: Continuity and change. Bloomington, IN: Indiana Consortium for Mental Health Services Research.
25. Regier DA, Narrow WE, Rae DS, Manderscheid RW, Locke BZ, Goodwin FK ( 1993) The de facto US mental and addictive disorders service system. Epidemiologic catchment area prospective 1-year prevalence rates of disorders and services. Arch Gen Psychiatry 50: 85-94. [Fulltext Link] [Medline Link]
26. Rosenstock IR ( 1966) Why people use health services. Milbank Memorial Fund Q 44: 94-106. [Medline Link]
27. Slovick P, Kraus NN, Lappe H, Letzel H, Malmfors T ( 1989) Risk perception of prescription drugs: Report on a survey in Sweden. In: Horisperger BR, Dinkel R (Eds), The perception and management of drug safety risks (pp 90-111). Berlin: Springer.
28. Streicker SK, Amdur M, Dincin J ( 1986) Educating patients about psychiatric medications: Failure to enhance compliance. Psychosoc Rehabil J 9: 15-28.
29. Swindle RS, Heller K, Pescosolido BA, Kikuzawa S ( 2000) Responses to "nervous breakdowns" in America over a 40-year period: Mental health policy implications. Am Psychol 55: 740-749. [Medline Link]
30. Thom DH, Ribisl KM, Stewart AL, Luke DA, Stanford Trust Study Physicians ( 1999) Further validation and reliability testing of the trust in physician scale. Med Care 37: 510-515. [Fulltext Link] [Medline Link] [CINAHL Link]
----------------------------------------------Accession Number: 00005053-200303000-00005 ----------------------------------------------