Anxiety diagnoses and their relationship to the number of prescribed psychotropic medications.

We designed this study to explore the relationship between an anxiety disorder diagnosis and the number of prescribed psychotropic medications. We retrospectively reviewed the medical records of 337 psychiatry outpatients, who were being seen in a psychotherapy training clinic, with regard to psychiatric diagnosis and the number of psychotropic medications. We grouped diagnoses into three categories: (a) anxiety disorders, (b) mood disorders, and (c) other Axis I disorders. Approximately 16% of persons were prescribed no psychotropic medications, while 41.5%, 30.0%, 8.6%, 2.7%, and 0.9% were prescribed one, two, three, four, and five psychotropic medications, respectively. The proportion of individuals on more than one medication with anxiety disorders was significantly greater than the base rate in the overall sample (z = 3.21, p < .001) or the proportion diagnosed with mood (z = 2.53, p = .01) or other Axis I disorders (z = 3.90, p < .001). In addition, for persons who were diagnosed with different diagnostic groupings, when one was anxiety disorders, they were more likely to be prescribed multiple psychotropic medications. In an outpatient psychiatry clinic, an anxiety disorder diagnosis appears to heighten the likelihood of being prescribed a greater number of psychotropic medications. We discuss the possible implications of these findings in outpatient mental health settings.

**********

Anxiety disorders are fairly prevalent in mental health settings (Kent & Gorman, 1997; Parker et al., 1999: Zimmerman, McDermut, & Mattia, 2000), often in combination with depression. For example, Parker and colleagues found that anxiety disorders were over-represented in a cohort of 269 individuals suffering from major depression, and Zimmerman and colleagues found that two-thirds of 373 depressed outpatients had a comorbid anxiety disorder.

With regard to the prescription of psychotropic medications among sufferers of anxiety disorders, there is scant empirical information. However, prescription patterns appear to be changing over time, both in the types as well as number of medications. For example, among a cohort of persons with anxiety disorders, Salzman, Goldenberg, Bruce, and Keller (2001) found that there were changes in prescribing patterns between the years 1989 and 1996, with exclusive benzodiazepine prescription shifting to a combination of benzodiazepines and an antidepressant (i.e., an increased number of medications). In examining the treatment of panic disorder and changes in prescribing patterns from 1992 to 1997, Uhlenhuth, Balter, Ban, Thomas, and Yang (1998) found an overall increase in the number of medications. Phillips and colleagues (2001) found that those anxiety sufferers with a comorbid personality disorder were prescribed a greater number of medications than those without Axis II psychopathology. Finally, Pillay (2002) clarified the reasons for polypharmacy among those with anxiety disorders, highlighting issues such as the potential for high levels of psychiatric comorbidity, current availability of broader types of medications, and clinician and individual medication patterns.

These preceding data suggest that anxiety sufferers may be prescribed more medications than those without anxiety, and that diagnostic comorbidity may be a contributory variable. However, previous studies have been limited by the absence of control groups (i.e., comparison of those with versus without anxiety disorders; Salzman et al., 2001); the examination of a specific anxiety disorder (e.g., panic disorder) rather than various types of anxiety disorders (Uhlenhuth et al., 1998); and the examination of comorbidity among anxiety sufferers with only one additional disorder (e.g., those with versus without personality disorder; Phillips et al., 2001).

We undertook this retrospective study among psychiatric outpatients to determine: (a) if an anxiety disorder diagnosis, versus other types of psychiatric disorders, increased the likelihood of being prescribed multiple psychotropic medications; and (b) whether Axis I comorbidity among those with anxiety disorders increased the likelihood of a greater number of prescriptions. In other words, compared with other Axis I disorders, did a diagnosis of anxiety disorder result in an overall higher number of prescribed medications? If so, was this relationship mediated, in part, by a greater number of additional psychiatric diagnoses (i.e., comorbidity)?

METHOD

Participants

Demographically, the study sample was made up of 218 women (65%), 77 men (23%), and 42 (13%) individuals whose sex was not noted in the medical record. The sample ranged in age from 18 to 80 years (M = 36.95, SD = 11.74). Approximately 48% of the sample was married, 26% single, 16% divorced, and 2% widowed; data were missing for 9%. Approximately 75% were Caucasian, though race could not be identified in 15% of the medical records. The highest level of education completed could not be identified for 34%. Of the remaining cases, only 56 (17% of the entire sample) had completed a college degree.

Procedure

In a resident-staffed, outpatient psychotherapy clinic of a university-based psychiatry-training program, a single investigator reviewed the medical records of all attendees between the years 1994 and 2001. Medical records were examined for demographic data, Axis I diagnoses, and the type (i.e., antidepressants, benzodiazepines, mood stabilizers, antipsychotics, other) as well as the number of prescribed psychotropic medications. The number of medications was tabulated at either the point of discharge for terminated persons, or at the time of the review for active attendees. That is, there was no standard point of medication assessment. The medication counts are a reflection of co-administered, not cumulative, prescriptions. All medications were prescribed by the psychotherapist/resident. Approximately 84% of the persons in this resident clinic were prescribed medications. In other words, most were exposed to both psychotherapy as well as adjunctive treatment with psychotropic medication.

We excluded records from review if: (a) individuals were under the age of 18; (b) the psychotherapy intervention was marital treatment (typically, there were not explicit psychiatric diagnoses noted for couples); (c) there was insufficient information to complete the data collection sheet or enable reasonable comparison with other subjects; and/or (d) the number of sessions was so few that psychiatric diagnoses were not confirmed. Of the 826 records examined, 337 (40.8%) met criteria for entry into this study.

Note that a sizeable number of medical records were excluded from review, most often because diagnoses were not available. In this particular psychotherapy clinic, psychiatric residents-in-training are advised not to confirm diagnoses in the medical record until the completion of two to four patient visits, which enabled supervisors to participate in diagnosis. Because the clinic does not accept insurances, expediency of diagnosis was not a relevant issue. The initial lack of diagnosis does not preclude indicated treatment, such as medication, psychoeducation, or acute problem-solving. However, being a clinic for the indigent, many persons do not attend a sufficient number of appointments to conclude a diagnosis.

To create reasonably sized study groups for comparison, we categorized diagnoses into three diagnostic groups: (a) anxiety disorders; (b) mood disorders; and (c) other Axis I disorders (e.g., eating disorders, substance-related disorders, somatoform disorders). Although we included all specific types of anxiety disorder diagnoses (e.g., panic disorder, social anxiety disorder, generalized anxiety disorder) into the anxiety disorder group, we specifically elected not to examine these diagnoses individually because the resulting study cells would be so small that statistical analysis would be compromised. After compilation, we compared diagnostic groupings with regard to being prescribed more than one psychotropic medication. We chose to examine the number of psychotropic medications as the study variable, rather than individual types of medications, because the latter approach would have resulted in a number of smaller subsamples and compromised meaningful statistical analysis.

RESULTS

Sample Diagnoses and Number of Psychotropic Medications Prescribed

We first examined in the entire sample the number patients diagnosed with anxiety disorders and the number of psychotropic medications that patients were prescribed. Of the 337 patients in the entire sample, 217 (64.4%) were not diagnosed with an anxiety disorder, while 107 (31.8%) were diagnosed with one, 12 (3.6%) with two, and 1 (0.3%) with three anxiety disorders. In the sample, 55 (16.3%) were prescribed no medication, 140 (41.5%) were prescribed one, and the remaining 142 (42.1%) were prescribed more than one medication: 101 (30.0% of the overall sample) were prescribed two, 29 (8.6%) three, 9 (2.7%) four, and 3 (0.9%) five psychotropic medications.

Patients with an Anxiety Disorder and the Likelihood of Being Prescribed Multiple Medications

We next examined whether being diagnosed with an anxiety disorder, regardless of comorbid diagnoses, would increase an individual's chances of being prescribed multiple medications. Of the 120 patients who were diagnosed with an anxiety disorder, 71 (59.2%) were prescribed multiple medications. Using z-tests, we found that this proportion was significantly greater than the base rate of being prescribed multiple medications in the overall sample (42.1%), z = 3.21, p < .001. Furthermore, this proportion was also significantly greater than the proportion of patients diagnosed with a mood disorder (45.2%, z = 2.53, p = .01) or another (i.e., non-anxiety, non-mood) Axis I disorder (33.3%, z = 3.90,p < .001).

From the overall sample, we also noted that having more than one Axis I diagnosis increased the chances that a patient would be prescribed multiple medications, as 29.5 % of the patients diagnosed with only one Axis I disorder were prescribed multiple medications compared with 50.2% of those who were diagnosed with more than one Axis I disorder, [chi square] (1) = 14.11, p < .001. Therefore, we next examined whether the number of diagnosed anxiety disorders had an effect on being prescribed multiple medications. Of the patients who were not diagnosed with an anxiety disorder, only 32.7% were prescribed multiple medications, whereas 55.1% of patients diagnosed with one and 92.3% of those diagnosed with more than one anxiety disorder were prescribed multiple medications, [chi square] (2) = 28.74, p < .001. Although this same pattern was found for mood disorders (31.1% with no mood disorders were prescribed multiple medications, compared with 43.3% with one and 56.4% with more than one mood disorder, [chi square] (2) = 7.09, p < .05), it was not as pronounced. As for other Axis I disorders, the pattern was not as clear: 45.9% of the patients without an other Axis I disorder were prescribed multiple medications compared with 29.4% of those diagnosed with one and 66.7% with more than one, [chi square] (2) = 8.58, p < .05.

The Number of Prescribed Psychotropic Medications and Comorbidity Among Diagnostic Groupings

Finally, because many persons in this sample were diagnosed with more than one type of Axis I disorder (i.e., anxiety, mood, other), we placed each patient into one of seven mutually exclusive categories, based on type of diagnosis and type of comorbidity. For example, a patient diagnosed with two disorders, both of which were anxiety disorders, was placed in the anxiety-disorders-only group, as were patients diagnosed with only one disorder that was an anxiety disorder. As can be seen in the Table, 50% or more of each diagnostic grouping that contains an anxiety disorder (anxiety disorders only, anxiety and mood, anxiety and other Axis I, and all three types) resulted in the prescription of multiple medications; whereas none of the other three groupings reached 50%.

DISCUSSION

The data indicate that, among psychiatric outpatients in a psychotherapy clinic, the likelihood of being prescribed more than one psychotropic medication is associated with the diagnosis of some type of anxiety disorder. In addition, among these individuals, comorbidity with either another anxiety or non-anxiety disorder diagnosis is associated with greater prescription of psychotropic medications. These findings suggest that anxiety disorder diagnoses have a predictable and consequential influence on the prescription profile for individuals.

The presence of an anxiety disorder diagnosis may indicate a more refractory individual (Crown et al., 2002), a phenomenon that may be intensified through the effects of comorbidity. Specifically, anxiety disorders are highly comorbid with other psychiatric disorders, most often depressive disorders (Brown, Campbell, Lehman, Grisham, & Mancill, 2001; Noyes, 2001). As an example, Bruce, Machan, Dyck, and Keller (2001) examined 179 clients with generalized anxiety disorder and found, at intake, that 39% had comorbid major depression. Four years after intake, the prevalence of comorbid major depression increased to 65%, and at 8 years to 74%. According to Noyes, comorbidity is associated with greater impairment, more treatment seeking, and a poorer prognosis.

It may also be that, as a result of our approach to diagnostic groupings, we have created "unequal" playing fields among the diagnoses, as there are potentially more anxiety disorder diagnoses (11 DSM diagnoses) in the anxiety disorder grouping than depressive diagnoses (9 DSM diagnoses) in the mood disorder grouping. However, the volume of individuals with mood disorders in this sample is substantially greater than those with anxiety disorders. In other words, given these diagnostic groupings, mood disorders remained far more prevalent than anxiety disorder diagnoses.

It is probable that, compared with other Axis I disorders, the treatment of anxiety disorders is associated with a broader range of potentially therapeutic psychotropic drugs. For example, in the treatment of anxiety disorders, Spiegel, Wiegel, Baker, and Greene (2000) emphasize the use of benzodiazepines and antidepressants; anxiety symptoms are an indication for buspirone; and Malt (2000) describes the use of anticonvulsants. In addition, some prescribers use antihistamines, beta blockers, and/or hypnotics (e.g., trazodone). In refractory cases, prescribers may add a second or third drug in an effort to augment the primary medication, including low-dose atypical antipsychotics. We did not specifically explore the different classes of medication in each diagnostic grouping, but one would expect that a greater number of medications would be associated with a greater number of different classes of medications (i.e., it is unlikely that prescribers would elect, for example, two medications that are antidepressants or two that are anticonvulsants).

Because this is a psychotherapy-training clinic, there were few persons in the database with diagnoses of schizophrenia or other psychoses. In addition, although various forms of psychotherapy are utilized in the clinic (e.g., insight-oriented psychotherapy, supportive psychotherapy, acute problem-solving, cognitive-behavioral approaches), there is no standardized psychotherapy approach. Our impression is that most trainees utilize a combination of techniques in addition to psychotropic medication. So, we were not able to control for the type of psychotherapy treatment received by attendees. Perhaps more effective psychotherapy treatments would reduce the need for adjunctive medications. Finally, during this study, there were no known pharmaceutical studies in progress in the clinic.

Implications for Mental Health Counselors 

From a clinical perspective, the combination of psychotherapy and psychotropic medication appears to enhance outcome, particularly when one or the other treatment alone is less robust (Thase, 2000). This general perspective has been affirmed for various anxiety disorders (Hollander & Simeon, 2003; Spiegel et al., 2000), including generalized anxiety disorder (Gorman, 2002; Rickels & Rynn, 2002). Thus, mental health counselors are likely to encounter anxious individuals who present for therapy while also taking medication. These data suggest that for the anxious individual who is seeking psychotherapy treatment, when medication is prescribed, there is likely to be more than one prescription. Multiple prescriptions may be the rule rather than the exception.

As with any type of intervention, the potential risks and benefits of medications need to be monitored. Due to the frequency of contacts with psychotherapy treatment, the mental health counselor is in the most advantageous position to determine if psychotherapeutic gains are being made in the treatment (i.e., whether or not the medications are facilitating the psychotherapy treatment). Ideally, the prescription of medications for the anxious individual in psychotherapy treatment is designed to assist with the modulation of anxiety and to enable a cognitive presence. Anxiety reduction is intended to facilitate the acquisition of skills and assist in the resolution of problem areas. The mental health counselor can also assess the risks of medication treatment, particularly medication side effects, possible medication interactions with substances such as alcohol and other prescription medications, and the potential risk for overdose. An additional risk is that some patients may perceive the medication as the answer, such that the prescription unintentionally sabotages psychotherapy treatment.

The mental health counselor can also encourage medication compliance. Medication compliance is a significant factor in treatment response. Compliance may be compromised by side effects, the cost of the medication, the dosing frequency of the medication, and resistance to adherence because of social stigma. With regard to side effects, many of the antidepressants (e.g., paroxetine, tricyclic antidepressants) are known to cause weight gain, which is an unacceptable side effect for many. As for cost, several medications are now available in generic form, but most of the newer antidepressants and anticonvulsants are not. Some medications require frequent dosing due to their short duration of action. As examples, buspirone must be dosed up to three times per day, and the anticonvulsant gabapentin is typically dosed two to three times per day. Clinically, more frequent dosing per day runs the risk of less compliance. Finally, there is still considerable social stigma with psychotropic medications. For some individuals, taking this type of medication labels them as mentally ill. As a result, some persons may be fearful of having others find out that they are prescribed such medications. Given these potential risks, the mental health counselor can, in addition to the prescriber, be invaluable in exploring concerns about medication and encouraging compliance. Many individuals simply need further reassurance about the anticipated therapeutic benefits, side effects, their fears (e.g., "Will this change my personality?"), and importance of compliance. In the counseling environment, these individuals may be more comfortable in expressing their concerns about unrelenting side effects, such as sexual dysfunction with selective serotonin reuptake inhibitors.

Limitations and Conclusions

There are a number of potential limitations with this study. First, all prescriptions were written by resident providers whose prescribing patterns may not mirror psychiatrists who have completed training; however, each resident had a faculty supervisor who guided both the psychotherapy and medication treatment of each individual. Second, this study took place in a psychotherapy clinic. Although we do not perceive any particular prescribing differences between this type of clinic and a medication clinic, differences may exist. Third, our diagnostic groupings may have unintentionally compromised the interpretation of data. However, our intention was to generate sufficiently sized study groups to reasonably compare prescribing trends. Fourth, we do not have data on specific medications because of potential difficulties with statistical analyses and, therefore, do not know if one or more particular medications explain these findings. Finally, we were only able to confirm the prescription of medication; prescription does not indicate compliance. So, whether or not these medications were actually utilized remains unknown.

Despite these limitations, to our knowledge, this is the first study to examine the impact of an anxiety disorder diagnosis on the number of psychotropic medications prescribed. These data indicate that those who suffer from anxiety are likely to be prescribed a greater number of medications than those without such symptoms. Future studies need to examine the deeper implications of these findings (i.e., does this represent treatment resistance, a broader range of potential medications in anxiety disorders, and/or the use of medications as augmentors?). In addition, these findings warrant verification in a different treatment setting, with a broader representation of race and gender. Future research may determine the explicit contributory factors for the relationship between anxiety disorder diagnoses and a greater number of prescribed psychotropic medications.

Number (%) of individuals prescribed more than one psychotropic
medication by Axis I diagnostic groupings

Axis I Diagnostic Group                      n   % Prescribed Multiple
                                                      Medications
Diagnosed with only one grouping
   Anxiety disorders, only                  22            50%
   Mood disorders, only                     37            32%
   "Other" Axis I disorders, only           47            19%
Diagnosed with more than one grouping
   Anxiety and mood disorders               65            62%
   Anxiety and "other Axis I" disorders      5            60%
   Mood and "other Axis I" disorders        55            46%
   Mood, anxiety, and "other Axis I"
      disorders                             28            61

Note: Sample N = 337.

REFERENCES

Brown, T. A., Campbell, L. A., Lehman, C. L., Grisham, J. R., & Mancill, R. B. (2001). Current and lifetime comorbidity of the DSM-IV anxiety and mood disorders in a large clinical sample. Journal of Abnormal Psychology, 110, 585-599.

Bruce, S. E., Machan, J. T., Dyck, I., & Keller, M. B. (2001). Infrequency of "pure" GAD: Impact of psychiatric comorbidity on clinical course. Depression and Anxiety, 14, 219-225.

Crown, W. H., Finkelstein, S., Berndt, E. R., Ernst, R., Ling, D., Poret, A. W., et al. (2002). The impact of treatment-resistant depression on health care utilization and costs. Journal of Clinical Psychiatry, 63, 963-971.

Gorman, J. M. (2002). Treatment of generalized anxiety disorder. Journal of Clinical Psychiatry, 63, 17-23.

Hollander, E., & Simeon, D. (2003). Concise guide to anxiety disorders. Washington, DC: Psychiatric Publishing.

Kent, J. M., & Gorman, J. M. (1997). Drug treatment of anxiety disorders with comorbidity. In S. Wetzler & W. C. Sanderson (Eds.), Treatment strategies for patients with psychiatric comorbidity (pp. 105-134). New York: Wiley.

Malt, U. F. (2000). The basis for the use of antiepileptic drugs in psychiatric disorders. Acta Neurologica Scandinavica Supplementum, 102, 40-42.

Noyes, R. (2001). Comorbidity in generalized anxiety disorder. Psychiatric Clinics of North America, 24, 41-55.

Parker, G., Wilhelm, K., Mitchell, P., Austin, M-P., Roussos, J., & Gladstone, G. (1999). The influence of anxiety as a risk to early onset depression. Journal of Affective Disorders, 52, 11-17.

Phillips, K. A., Shea, M. T., Warshaw, M., Dyck, I., Bruce, S., & Keller, M. (2001). The relationship between comorbid personality disorders and treatment received in patients with anxiety disorders. Journal of Personality Disorders, 15, 157-167.

Pillay, S. S. (2002). Polypharmacy of anxiety disorders. In S. N. Ghaemi (Ed.), Polypharmacy in psychiatry (pp. 133-150). New York: Marcel Dekker.

Rickels, K., & Rynn, M. (2002). Pharmacotherapy of generalized anxiety disorder. Journal of Clinical Psychiatry, 63, 9-16.

Salzman, C., Goldenberg, I., Bruce, S. E., & Keller, M. B. (2001). Pharmacologic treatment of anxiety disorders in 1989 versus 1996: Results from the Harvard/Brown Anxiety Disorders Research Program. Journal of Clinical Psychiatry, 62, 149-152.

Spiegel, D. A., Wiegel, M., Baker, S. L., & Greene, K. A. I. (2000). Pharmacological management of anxiety disorders. In D. I. Mostofsky & D. H. Barlow (Eds.), The management of stress and anxiety in medical disorders (pp. 36-65). Boston: Allyn & Bacon.

Thase, M. E. (2000). Psychopharmacology in conjunction with psychotherapy. In C. R. Snyder & R. E. Ingram (Eds.), Handbook of psychological change: Psychotherapy processes and practices for the 21st century (pp. 474-497). New York: John Wiley.

Uhlenhuth, E. H., Balter, M. B., Ban, T. A., Thomas, A., & Yang, K. (1998). International study of expert judgment on therapeutic use of benzodiazepines and other psychotherapeutic medications: V. Treatment strategies in panic disorder, 1992-1997. Journal of Clinical Psychopharmacology, 18, 27S-31S.

Zimmerman, M., McDermut, W., & Mattia, J. I. (2000). Frequency of anxiety disorders in psychiatric outpatients with major depressive disorder. American Journal of Psychiatry, 157, 1337-1340.

Randy A. Sansone, M.D., is a professor in the Departments of Psychiatry and Internal Medicine at Wright State University School of Medicine in Dayton, OH, and Director of Psychiatry Education at Kettering Medical Center in Kettering, OH. E-mail: [e-mail] George A. Gaither, Ph.D., is an assistant professor in the Department of Psychological Science at Ball State University in Muncie, IN. Dorota Rytwinski, M.D., is a senior resident in the Department of Psychiatry at Wright State University School of Medicine in Dayton, OH.