Assessment and Diagnosis “Under the Gun”

The Influence of Payment Method on Psychologists’ Diagnostic Decisions Regarding Minimally Impaired Clients

Andrew M. Pomerantz and Dan J. Segrist Department of Psychology

Southern Illinois University Edwardsville

Are psychotherapy clients who pay via health insurance more likely to receive Diag- nostic and Statistical Manual of Mental Disorders (4th ed. [DSM–IV], American Psychiatric Association, 1994) diagnoses than identical clients who pay out of pocket? Previous research (Kielbasa, Pomerantz, Krohn, & Sullivan, 2004) indicates that when psychologists consider a mildly depressed or anxious client, payment method significantly influences diagnostic decisions. This study extends the scope of the previous study to include clients whose symptoms are even less severe. Independ- ent practitioners responded to vignettes of clients whose profiles deliberately in- cluded subclinical impairment and a high level of functioning. Half of the partici- pants were told that the clients would pay via managed care; the other half were told that the clients would pay out of pocket. As in the earlier study, payment method had a highly significant impact on diagnosis such that relative to out-of-pocket clients, managed care clients were much more likely to be assigned DSM–IV diagnoses. In addition, a noteworthy percentage of participants assigned diagnoses regardless of payment method. Ethical implications are discussed.

Keywords: diagnosis, managed care, independent psychotherapy practice, ethics, payment

A recent study (Kielbasa, Pomerantz, Krohn, & Sullivan, 2004) found that the method by which a private practice client pays for psychotherapy very signifi- cantly influences both the likelihood that the psychologist will assign a diagnosis and the specific diagnosis that the psychologist chooses. Specifically, when com-

ETHICS & BEHAVIOR, 16(3), 253–263 Copyright © 2006, Lawrence Erlbaum Associates, Inc.

Correspondence should be addressed to Andrew M. Pomerantz, Southern Illinois University Edwardsville, Department of Psychology, Edwardsville IL 62026. E-mail: [email protected]

pared to identical clients paying out of pocket, clients paying via managed care were much more likely to receive diagnoses and were more likely to receive adjust- ment disorder diagnoses in particular. In the Kielbasa et al. (2004) study, the fic- tional vignettes to which participants responded included mildly depressed and anxious clients whose symptoms placed them very near the threshold for Axis 1 disorders in the Diagnostic and Statistical Manual of Mental Disorders (4th ed. [DSM–IV], American Psychiatric Association [APA], 1994). The purpose of this study was to replicate the Kielbasa et al. (2004) study using vignettes of clients whose level of pathology was even less severe. In other words, when considering a generally high functioning client whose symptoms may fall below the threshold for any DSM–IV disorder, will psychologists be influenced by payment method when making diagnostic decisions?

This study, as well as its predecessor (Kielbasa et al., 2004), stems from a grow- ing body of literature examining the effects of managed care and other forms of third-party payment on the independent practice of psychology. This literature in- cludes numerous empirical surveys of practitioners regarding the impact of man- aged care on their practices (e.g., Bell, 1999; Murphy, DeBernardo, & Shoemaker, 1998; Phelps, Eisman, & Kohout, 1998; Rothbaum, Bernstein, Haller, Phelps, & Kohout, 1998), most of which have concluded that the impact has been quite nega- tive. In addition, the literature includes nonempirical commentaries on the impact of managed care on psychotherapy (e.g., Karon, 1995; Miller, 1996), most of which describe managed care as having a detrimental effect. However, diagnosis has not been the primary focus of these surveys and commentaries. Instead, they have focused primarily on the therapy process, and to a much more limited extent, assessment techniques. Very few studies in this field have yielded conclusions re- garding diagnostic issues; these studies have found that independent practitioners strongly believe that managed care influences psychologists to alter diagnoses to ensure reimbursement and to protect confidentiality (Murphy et al., 1998) and that accurate diagnosis in a managed care system is problematic for many mental health counselors (Danziger & Welfel, 2001). Thus, with the exception of these isolated studies and the Kielbasa et al. (2004) article, the impact of payment method on specific diagnostic decisions made by clinicians has not been the focus of empirical research, particularly for clients who present with mild or sub- threshold symptoms.

Kielbasa et al. (2004) considered numerous interpretations for the finding that managed care clients are far more likely to receive diagnoses than out-of-pocket clients, including the common requirement of a diagnosis by managed care com- panies to justify reimbursement as described by Ackley (1997), Chambliss (2000), and Kutchins and Kirk (1997), among others. Peck and Scheffler (2002) similarly discuss “intentional upcoding,” by which clinicians exaggerate symptoms to in- crease the chances or amount of reimbursement from third-party payers (p. 1094). To the extent that the client falls below the criteria for a mental disorder, the clini-

254 POMERANTZ AND SEGRIST

cian engaging in such practices may be acting in a manner that violates the ethical code of the APA (2002) or laws concerning insurance fraud. In light of this, this study specifically sought to examine the impact of payment method on diagnostic decisions regarding clients whose symptoms may fall below the threshold of a DSM–IV disorder.

METHOD

Participants

Members of Division 42 (Psychologists in Independent Practice) of the APA were randomly selected and surveyed via mail. Of the 1,000 members who were initially surveyed, 91 surveys were returned as undeliverable. Of the remaining 909 surveys that were presumably delivered, 275 respondents provided usable data, represent- ing a 30.25% return rate. Mean age of participants was 59.4 years (SD = 9.63) and mean number of years in private practice was 23.8 (SD = 10.11). Respondents were primarily male (63%) and White (78%). Most had earned PhD degrees (86%) as opposed to EdD (6%) or PsyD (7%) degrees, and most specialized in clinical psychology (74%) as opposed to counseling psychology (20%) or other areas (4%). Eclectic orientation was most frequently endorsed (57%), followed by cog- nitive (18%) and psychodynamic (16%). Most (62%) worked primarily in solo in- dependent practices, whereas some (22%) worked primarily in group independent practices.

Materials, Design, and Procedure

Each participant received a survey that included two vignettes, one describing a cli- ent with minimal depressive symptoms and another describing a client with minimal anxious symptoms. The order of the two vignettes was counterbalanced such that each appeared first on approximately half of the surveys. For participants in the man- aged care condition, both fictional clients were described as paying via managed care. For participants in the out-of-pocket condition, both fictional clients were de- scribed as paying out of pocket. The Appendix provides an illustration of these vi- gnettes. Participants also received a cover letter and a brief demographic survey.

Each vignette was intended to portray a client whose presenting problems fell below the threshold for a DSM-IV diagnosis. They were similar to the vignettes used in the Kielbasa et al. (2004) study in terms of the type of symptoms described, but the severity and duration of the symptoms were deliberately subclinical, and the level of functioning was deliberately high. Specifically, the minimally anxious client is described as demonstrating some symptoms of generalized anxiety disor- der, but they are insufficient in number and duration (only a “couple of months”

PAYMENT METHOD AND DIAGNOSIS 255

rather than 6 months) to qualify for this diagnosis. Similarly, the minimally depres- sive client is described as demonstrating some symptoms of a major depressive ep- isode, but they are insufficient in number and duration (only a week rather than 2 weeks) to qualify for this diagnosis. In both vignettes, the client was described as “generally functioning well in all areas of [his/her] life,” and experiencing symp- toms that have “typically passed quickly and have caused only slight impairment.” These two phrases closely mimic the language found in the “71–80” and “81–90” range of the Global Assessment of Functioning scale of the DSM-IV. They were in- tentionally incorporated into the vignettes to corroborate that, as the minimal symptoms imply, these clients are subclinically impaired and are functioning at a relatively high level. In addition, the vignettes included the statements that the cli- ents can identify “no specific triggers” for their symptoms; this statement was in- cluded as an attempt to ensure that the clients could not qualify for adjustment dis- order diagnoses.

After reading each vignette, which included presenting problem, symptoms, and some background and demographic information, participants responded to four questions: (a) “Would you assign this client a DSM–IV diagnosis?” (“yes” and “no” choices provided), (b) “If you answered yes to the previous question, what specific diagnosis would you provide?” (blank space, rather than specific choices, provided), (c) “Assuming that the client does not prematurely terminate, predict the length of therapy, in number of sessions,” and (d) “What prognosis would you give this client?” For the final question, participants were provided with a 5-point Likert-type scale ranging from 1 (extremely poor) to 5 (extremely good).

RESULTS

Tables 1 and 2 display the frequencies of “yes” and “no” responses, for both payment conditions, to the item, “Would you assign this client a DSM–IV diagnosis?” As the tables illustrate, assignment of a diagnosis was more common for managed care cli- ents than for out-of-pocket clients across both vignettes. The percentage of “yes” and “no” responses to the managed care condition were used as comparisons for the out-of-pocket condition in two chi-square tests for goodness of fit (one for each vi- gnette). For the first vignette (minimally anxious symptoms), χ2(1, N = 270) = 22.29 p < .001, and for the second vignette (minimally depressive symptoms, χ2 (1, N = 272) = 21.02, p < .001, there was a significant association between the method of payment and whether or not the client would be diagnosed. These highly significant chi-square statistics indicate that the likelihood of a participant assigning a diagnosis to a minimally impaired client paying via managed care is significantly higher than the likelihood of a participant assigning a diagnosis to the same client paying out of pocket. Specifically, based on the odds ratio calculated from data in all four cells of the chi-square, the client with minimally depressive symptoms paying via managed

256 POMERANTZ AND SEGRIST

care was 3.17 times more likely than an identical client paying out of pocket to be di- agnosed with a DSM disorder. Similarly, based on the odds ratio, the minimally anx- ious client paying with managed care was 3.33 times more likely to be diagnosed with a DSM disorder than an identical client paying out of pocket.

Anecdotally, it is noteworthy that 9.68% of participants in the managed care condition who responded “yes” to the initial question (“Would you assign this cli- ent a DSM–IV diagnosis?”) included unsolicited written comments about third- party reimbursement. Specifically, these participants inserted comments to the ef- fect that assigning a diagnosis was a necessity for payment or reimbursement. Ex- amples of these comments include, “She has not had symptoms long enough to re- ally warrant the diagnosis but I would need to give her a diagnosis to bill her HMO,” and “Need a code for insurance.” The appearance of these comments is es- pecially interesting because there was no request or designated space for com- ments on the questionnaire.

The item “If you answered yes to the previous question, what specific diagnosis would you provide?” generated a wide variety of responses from participants. Ta-

PAYMENT METHOD AND DIAGNOSIS 257

TABLE 1 Frequencies of “Yes” and “No” Responses to the Item “Would You Assign

This Client a DSM-IV Diagnosis?” by Payment Method Regarding Vignette No. 1 (Minimally Anxious Client)

Yes No

Frequency % Frequency % Total

Managed care 90 72.0 35 28.0 125 Out of pocket 63 43.4 82 56.6 145 Total 153 117 270

Note. DSM-IV = Diagnostic and Statistical Manual of Mental Disorders (4th ed.; American Psy- chiatric Association, 1994).

TABLE 2 Frequencies of “Yes” and “No” Responses to the Item “Would You Assign

This Client a DSM-IV Diagnosis?” by Payment Method Regarding Vignette No. 2 (Minimally Depressive Client)

Yes No

Frequency % Frequency % Total

Managed care 79 63.7 45 36.3 124 Out of pocket 53 35.8 95 64.2 148 Total 132 140 272

Note. DSM-IV = Diagnostic and Statistical Manual of Mental Disorders (4th ed.; American Psy- chiatric Association, 1994).

bles 3 and 4 display the frequencies and percentages of the specific diagnoses pro- vided for each vignette.

Regarding estimated number of sessions needed to treat each client, partici- pants predicted a mean of 9.97 sessions (SD = 9.92) for the minimally anxious client, with a mean prognosis of 4.40 (SD = .65). For the client with minimally depressive symptoms participants predicated a mean of 10.08 sessions (SD = 10.79) and a prognosis of 4.43 (SD = .62). Independent samples t tests indi- cated that portrayed method of payment had no significant effect on estimates of the duration of therapy or the prognosis of the hypothetical clients. Spe- cifically, there was no significant difference in the estimated duration of ther- apy for the minimally anxious client paying for therapy through managed care, M = 9.43, SD = 8.04, and the same client paying for therapy out of pocket, M = 10.38, SD = 11.16; t(229) = –.72, p = .47. In addition, there was no significant difference in the perceived prognosis of the minimally anxious client paying through managed care, M = 4.35, SD = .71, and paying out of pocket, M = 4.45, SD = .60; t(221.36) = –1.22 , p = .23. Similarly there was no significant difference in the estimated duration of therapy for the client with minimally depressive symp- toms paying for therapy through managed care, M = 10, SD = 9.76, and the same client paying for therapy out of pocket, M = 10.15, SD = 11.52; t(228) = –.13, p = .90. There was also no significant difference in the perceived prognosis of the cli- ent with minimally depressive symptoms paying through managed care, M = 4.41, SD = .65, and paying out of pocket, M = 4.45, SD = .60; t(256) = –.58, p = .57.

258 POMERANTZ AND SEGRIST

TABLE 3 Frequencies and Percentages of DSM-IV-TR Diagnoses Assigned

to Vignette No. 1 (Minimally Anxious)

DSM-IV Diagnosis Frequency %

Adjustment disorder with anxiety 46 31.1 Anxiety NOS 41 27.7 Generalized anxiety disorder 22 14.9 Adjustment disorder with mixed anxiety and depression 20 13.5 Adjustment disorder NOS 5 3.4 Diagnosis deferred 5 3.4 Depression NOS 4 2.7 Dysthymic disorder 2 1.4 Major depression, recurrent, moderate 1 .7 Panic disorder with agoraphobia 1 .7 V code 1 .7

Note. DSM-IV = Diagnostic and Statistical Manual of Mental Disorders (4th ed.; American Psy- chiatric Association, 1994); DSM-IV-TR = DSM-IV (Text Rev., 2000); NOS = not otherwise specified. Sample (n = 148) composed of those participants who responded “Yes” to the question, “Would you as- sign this client a DSM-IV diagnosis?”

It is also notable that, regardless of payment method, a sizable number of partic- ipants chose to assign diagnoses to clients whose symptoms and level of function- ing suggested that no diagnosis was applicable. Altogether, 56% and 48% of the participants assigned a diagnosis to the minimally anxious client and minimally depressive client, respectively.

DISCUSSION

Results of this study suggest that, even when the client seeking services does not meet criteria for a mental disorder, the method by which the client pays for psycho- logical services has a very strong influence on the diagnostic decisions of the psy- chologist providing the services. Specifically, these results indicate that, relative to identical clients who pay out of pocket, clients who pay via managed care are far more likely to be diagnosed with a DSM–IV disorder. This finding was previously established with fictional clients whose symptoms were more serious (Kielbasa et al., 2004), but results of this study confirm that the finding holds true even when the clients’ issues fall below diagnosable levels. The finding that clients paying via managed care are more likely to receive diagnoses than those paying out of pocket is illustrated not only by the significant chi-square statistics reported earlier, but also by the odds ratio, which indicates that clients paying via managed care are

PAYMENT METHOD AND DIAGNOSIS 259

TABLE 4 Frequencies (Percentages) of DSM-IV-TR Diagnoses Assigned to Vignette

No. 2 (Minimally Depressive)

DSM-IV-TR Diagnosis Frequency %

Adjustment disorder with depression 38 29.7 Depression NOS 37 28.9 Dysthymic disorder 20 15.6 Adjustment disorder with mixed anxiety and depression 10 7.8 Diagnosis deferred 10 7.8 Adjustment disorder NOS 5 3.9 Major depression, single episode, mild 3 2.3 Adjustment disorder mixed disturbance of emotion and conduct 1 .8 Anxiety NOS 1 .8 Generalized anxiety disorder 1 .8 Major depression, recurrent, moderate 1 .8 V code 1 .8

Note. DSM-IV = Diagnostic and Statistical Manual of Mental Disorders (4th ed.; American Psy- chiatric Association, 1994); DSM-IV-TR = DSM-IV (Text Rev., 2000); NOS = not otherwise specified. Sample (n = 128) composed of those participants who responded “Yes” to the question, “Would you as- sign this client a DSM-IV diagnosis”?

slightly over three times more likely to be assigned a DSM–IV diagnosis than iden- tical clients paying out of pocket.

Kielbasa et al. (2004) offered the interpretation that the policy of many man- aged care companies to require a diagnosis for payment (Ackley, 1997; Chambliss, 2000; Kutchins & Kirk, 1997) was a likely reason for the finding that payment method influences diagnostic decision making. The same interpretation continues to apply to these results. In fact, these results (both empirical and anecdotal) sug- gest that psychologists may upcode clients whose symptoms are subclinical to se- cure payment when those clients pay via managed care but not when they pay out of pocket. Kielbasa et al. also suggested that diagnosing a client without merit can have various negative consequences: an unwarranted view of the self as mentally ill (Ackley, 1997; Caplan, 1995); the presence of a mental disorder on the client’s medical record, the confidentiality of which is not entirely in the hands of the clini- cian (Kutchins & Kirk, 1997); or an altered course of treatment (Beutler & Har- wood, 2000; Nathan & Gorman, 2002). All of these negative consequences are es- pecially applicable to a situation in which a client receives an unwarranted diagnosis. In addition, Kielbasa et al. described the importance of this finding for the clinicians themselves, whose behavior may be unethical (APA, 2002) or illegal (per insurance fraud regulations), and for society at large, which will pay higher in- surance rates to cover unwarranted claims. Both of these outcomes are particularly salient in the context of the findings of this study, which extend the findings of the Kielbasa et al. study to subclinical clients.

Although the primary intent of this study was to compare responses to managed care and out-of-pocket clients, it is striking that, regardless of payment method, ap- proximately half of the participants chose to assign diagnoses to these subclinical clients. This result suggests that clinicians may tend to overestimate clients’ symp- toms or apply DSM–IV criteria quite liberally in all payment circumstances, and that the presence of a managed care party in the therapy situation simply exacer- bates this tendency. The tendency is troubling, and its causes are not entirely clear. Perhaps psychologists themselves generally view diagnosis as a prerequisite for treatment; perhaps the increasing presence and influence of managed care in the professional lives of psychologists (e.g., Murphy et al., 1998) has contributed to this view. In any case, this issue is worthy of further study.

As in the Kielbasa et al. (2004) study, the importance of informed consent is a primary ethical implication of the results of this study. The APA ethics code (2002) clearly emphasizes the necessity of informed consent for clinical services (e.g., standards 3.10 [a], 10.01 [a]). Numerous authors have debated the specific content to be included in an adequate informed consent procedure (e.g., Appel- baum, 1993; Newman & Bricklin, 1991; Pope & Vasquez, 1998), but it seems reasonable to consider the effect of payment method on diagnostic decisions among the essential points to include, especially if the client’s symptoms are as minimal as those included in this study. These and previous (e.g., Kielbasa et al.,

260 POMERANTZ AND SEGRIST

2004) findings suggest that, to obtain truly informed consent, psychologists should advise clients that their method of payment might influence the diagnos- tic decisions made about them.

This study focuses only on what psychologists decide in terms of diagnosis. One potential avenue for future studies is an exploration of how psychologists communicate their diagnostic decisions to clients. The impact of the diagnosis on clients may vary greatly as a result of the manner and extent to which psycholo- gists share, explain, or discuss it with them.

Several limitations apply to the results and conclusions of this study. It incorpo- rated brief vignettes with limited information rather than more detailed vignettes or actual clients. The fictional clients were described as presenting with minimal symptoms and as relatively high functioning; it is unclear how common such subclinical presentations are in the practices of most independent practice psy- chologists. Respondents included a subset of the membership of Division 42, and the extent to which respondents are representative of the division or the profession as a whole is unclear. In spite of these limitations, the results of this study confirm and extend the results of previous research (Kielbasa et al., 2004) demonstrating the importance of payment method in the diagnostic decisions of psychologists in independent practice.

ACKNOWLEDGMENTS

This research was conducted with the support of a Dean’s Grant for Research Enhancement from the School of Education at Southern Illinois University Edwardsville.

REFERENCES

Ackley, D. C. (1997). Breaking free of managed care. New York: Guilford. American Psychiatric Association. (1994). The diagnostic and statistical manual of disorders (4th ed.).

Washington, DC: Author. American Psychological Association. (2002). Ethical principles of psychologists and code of conduct.

American Psychologist, 57, 1060–1073. Appelbaum, P. S. (1993). Legal liability and managed care. American Psychologist, 48, 251–257 Bell, P. J. (1999). Changes in the therapeutic process integrity. In K. Weisbarger (Ed.), The traumatic

bond between the psychotherapist and managed care (pp. 87–101). Northvale, NJ: Jason Aronson. Beutler, L. E., & Harwood, T. M. (2000). Prescriptive psychotherapy: A practical guide to systematic

treatment selection. New York: Oxford University Press. Caplan, P. J. (1995). They say you’re crazy: How the world’s most powerful psychiatrists decide who’s

normal. Reading, MA: Addison-Wesley. Chambliss, C. H. (2000). Psychotherapy and managed care. Needham Heights, MA: Allyn and Bacon.

PAYMENT METHOD AND DIAGNOSIS 261

Danziger, P. R., & Welfel, E. R. (2001). The impact of managed care on mental health counselors: A survey of perceptions, practices, and compliance with ethical standards. Journal of Mental Health Counseling, 23, 137–150.

Karon, B. P. (1995). Provision of psychotherapy under managed health care: A growing crisis and na- tional nightmare. Professional Psychology: Research and Practice, 19, 223–225.

Kielbasa, A. M., Pomerantz, A. M., Krohn, E. J., & Sullivan, B. F. (2004). How does clients’ method of payment influence psychologists’ diagnostic decisions? Ethics & Behavior, 14, 187–195.

Kutchins, H., & Kirk, S. A. (1997). Making us crazy: DSM: The psychiatric bible and the creation of mental disorders. New York: Free Press.

Miller, I. J. (1996). Managed care is harmful to outpatient mental health services: A call for account- ability. Professional Psychology: Research and Practice, 27, 349–363.

Murphy, M. J., DeBernardo, C. R., & Shoemaker, W. E. (1998). Impact of managed care on independ- ent practice and professional ethics: A survey of independent practitioners. Professional Psychol- ogy: Research and Practice, 29, 43–51.

Nathan, P. E., & Gorman, J. M. (Eds.). (2002). A guide to treatments that work (2nd ed.). New York: Oxford University Press.

Newman, R., & Bricklin, P. M. (1991). Parameters of managed mental health care: Legal, ethical, and professional guidelines. Professional Psychology: Research and Practice, 22, 26–35.

Peck, M. C., & Scheffler, R. M. (2002). An analysis of the definitions of mental illness used in state par- ity laws. Psychiatric Services, 53, 1089–1095.

Phelps, R., Eisman, E. J., & Kohout, J. (1998). Psychological practice and managed care: Results of the CAPP practitioner study. Professional Psychology: Research and Practice, 29, 31–36.

Pope, K. S., & Vasquez, M. J. T. (1998). Ethics in psychotherapy and counseling (2nd ed.). San Fran- cisco: Jossey-Bass.

Rothbaum, P. A., Bernstein, D. M., Haller, O., Phelps, R., & Kohout, J. (1998). New Jersey psycholo- gists’ report on managed mental health care. Professional Psychology: Research and Practice, 29, 37–42.

APPENDIX Sample Vignettes

Vignette #1

A 35-year-old married man in good physical health is a father of two children. He generally functions well in all areas of his life; however, for the past couple of months he has been experiencing increased anxiety and worry. He cannot identify any stressful events that may have triggered this worry. Specifically, he has experi- enced mild anxiety about work; temporary difficulty concentrating; and occa- sional difficulty sleeping. These symptoms have typically passed quickly and have caused only slight impairment.

This person has come to you today seeking therapy. He intends to pay for ther- apy [through his managed care plan/out of pocket].

Vignette #2

A 55-year-old married woman in good physical health is a mother of two children. She generally functions well in all areas of her life; however, for the past week she

262 POMERANTZ AND SEGRIST

has been experiencing increased sadness. She cannot identify any stressful events that may have triggered this sadness. Specifically, she has experienced a depressed mood; decreased interest in her activities and hobbies; and some difficulty sleep- ing. These symptoms typically last for a short period of time, and have caused only slight impairment.

This person has come to you today seeking therapy. She intends to pay for ther- apy [through her managed care plan/out of pocket].

PAYMENT METHOD AND DIAGNOSIS 263