Psychosomatics 41:512-518, December 2000
© 2000 The Academy of Psychosomatic Medicine
Psychiatric Symptoms and Medical Utilization in Primary Care Patients
Lisa A. Carbone, M.D.,
Arthur J. Barsky, M.D.,
E. John Orav, Ph.D.,
Alison Fife, M.D.,
Gregory L. Fricchione, M.D.,
Sarah L. Minden, M.D., and
Jonathan F. Borus, M.D.
Received December 28, 1999; revised March 16, 2000; accepted June 19, 2000. From Brigham and Women's Hospital, Boston, Massachusetts. Address correspondence and reprint requests to Dr. Carbone, M.D., Brigham and Women's Hospital, 221 Longwood Avenue, Boston, MA 02115.
|
ABSTRACT
|
In two studies, the authors evaluated the impact of psychiatric disorders on medical care utilization in a primary care setting. In the first study, 526 consecutive patients in a teaching hospital primary care practice completed the 18-item RAND Mental Health Inventory to identify clinically significant depression and/or anxiety and a questionnaire about the use of psychiatric treatment and psychoactive medications. The medical utilization of those patients defined as depressed and/ or anxious was compared with those defined as not depressed and/or anxious. Patients identified as depressed and/or anxious reported significantly increased medical utilization, but this was not confirmed by the hospital's computerized record system. In the second study, the authors analyzed medical care utilization for the years before and after the first outpatient psychiatry appointment of a sample of 91 patients referred from the same primary care practice to the hospital's outpatient psychiatry clinic over a 1-year period. In both studies there was not a statistically significant difference in medical utilization among those patients receiving psychiatric treatment. The findings demonstrate the difficulties in examining cost offset in a primary care population and raise questions about it as a realistic outcome measure of the effect of psychiatric treatment.
Key Words: Primary Care • Medical utilization • Psychiatric symptoms
|
INTRODUCTION
|
Psychiatric disorders are prevalent in primary care populations, with approximately 25% of all adult primary care patients having a psychiatric disorder and an additional 12.5% having subthreshold conditions.1–4 These disorders are frequently untreated, are especially prevalent among the highest utilizers of primary care services, and are associated with disproportionately high general health care costs.5–9 Katon et al.9 found that high medical utilizers had elevated levels of DSM-III-R disorders (e.g., major depression 23.5%, dysthymic disorder 16.8%, generalized anxiety disorder 21.8%, and somatization disorder 20.%) and that 67% of these patients were undertreated psychiatrically. It has been posited that treating the psychiatric disorders of primary care patients might decrease medical utilization and thereby have an offset effect on patients' total health care costs.
A variety of researchers have examined whether outpatient psychiatric treatment results in a decrease in general medical utilization. For example, a 1965 study in West Germany showed a decreased use of medical hospital care over a 5-year period for patients in psychoanalysis or psychoanalytic psychotherapy when compared with a control group.10 Subsequent studies have attempted to demonstrate a reduction in the overall health care costs of patients receiving psychiatric treatment; although some studies have shown reduced costs from some types of psychiatric treatment,11–17 the cost-offset effect has been most convincing within inpatient settings.14,15 In the outpatient setting it is clear that psychiatric disorders increase costs; however, data proving that outpatient treatment of psychologic distress will offset cost by decreasing medical utilization is more elusive.5,18–21
We undertook two studies to evaluate the medical utilization of psychiatric patients in a primary care setting. In our first study, we screen consecutive primary care patients for psychiatric symptoms, mental health treatment, psychotropic drug use, and medical utilization in the past year in order to assess the relationship between psychiatric symptoms, psychiatric care, and medical care. In particular, we compare the medical utilization of those with and without psychiatric symptoms and those receiving or not receiving psychiatric care. In our second study, we examine if patients referred to the psychiatry clinic by the same primary care setting show a decrease in their medical utilization in the year after the referral as compared with the prior year and if the amount of psychiatric treatment received was related to medical utilization.
|
METHODS
|
Setting
The Brigham and Women's Hospital (BWH) is a 735- bed teaching hospital with a large, on-site primary care practice, Brigham Internal Medicine Associates (BIMA), and an Outpatient Psychiatry Clinic (OPC) in a separate hospital building is one block away.
Subjects and Procedure
In our first study, 526 consecutive patients coming to BIMA for a scheduled medical appointment, on randomly selected days, were screened at the time of their visit.
In our second study, we reviewed referrals to the OPC from January 1996 to January 1997 in order to identify those patients referred from BIMA. From the total referrals tracked over the year, we did not include those patients that had no prior medical utilization, leaving 91 patients that could be evaluated. We divided this group into three cohorts: Group A patients were referred but never attended the OPC; Group B patients made three or fewer visits to the OPC in the year following the referral; and Group C patients came to the OPC for more than three visits in the succeeding year. The Brigham and Women's Hospital Human Research Committee approved both studies.
Measures
In our first study, we assessed psychiatric symptoms with the 18-item RAND Mental Health Inventory (MHI).22,23 This instrument has been widely used in general medical settings and has shown acceptable psychometric properties.22,23 The MHI comprises four subscales: anxiety, depression, emotional control, and positive well being. In this study we used only the anxiety and depression subscales, which consist of four items each. We assessed previous medical care utilization in our first study by a questionnaire and by searching the hospital's computerized records. Patients completed a self-report questionnaire indicating how many times they had been admitted to a hospital, visited an emergency room (ER), visited a clinic or doctor outside the BWH, or visited a mental health professional in the past year. In addition, we asked patients if they had taken any medication for depression, anxiety, or other emotional difficulty in the past year. The self-report questionnaire was developed for this purpose and has face validity.
We obtained utilization data within the BWH system for both studies from the hospital's computerized (automated) database. For our first study, we obtained total inpatient days, outpatient visits, and total costs for the 12 months preceding and the 8 months following the completion of the screening questionnaire in BIMA. For our second study, we obtained the same variables for the year before and after the referral date from BIMA to the OPC.
The BWH's Computer Assisted Hospitalization Analysis for the Study of Efficacy (CHASE) is a repository of databases including demographic, clinical, financial, and utilization data. The BWH-CHASE system records all inpatient stays and all outpatient visits within the BWH, and it can also provide total costs for both inpatient and outpatient care.
Data Analysis
For the first study, for interpretability, we dichotomized the MHI depression and anxiety scores using the bottom quartile of our sample to define depression (score 15) and anxiety (score14). We then classified patients as being depressed and/or anxious if they fell below the threshold for either scale. We compared cost and utilization measures between the two groups using the Wilcoxon rank sum test because our measures of utilization (i.e., costs, admissions, ER visits, and outpatient visits) were not normally distributed. Other subgroup comparisons were also carried out using the Wilcoxon rank sum test. We considered a two-sided P-value 0.05 as statistically significant. Despite the use of nonparametric tests, we used means to summarize our measures of utilization because, with relatively rare events such as admissions and ER visits, the medians end up being 0 in both groups and are uninformative.
We also researched the robustness of our findings by applying more stringent criteria to our definition of depression/anxiety. We used the bottom decile of depression (score11) and/or anxiety (score10) to define depression/anxiety, and then we repeated all comparisons. The result remained the same as above and are not reported further.
For our second study, we divided patients into three groups according to whether they failed to show up for any psychiatric counseling, attended one to three sessions, or attended more than three sessions. We compared cost and utilization data between these groups using the Kruskal- Wallis test.
|
RESULTS
|
In the first study, 526 patients were screened in the primary care clinic (71% women, 29% men; average age= 44). Of these patients, 174 (33%) were defined as depressed and/ or anxious and 352 (67%) were defined as neither depressed nor anxious based on our cut-offs for their scores on the RAND MHI.
In the year before screening, patients defined as depressed and/or anxious reported significantly more hospitalizations, ER visits, and other nonmental health visits than those patients who were neither depressed nor anxious (Table 1). However, for the same time period, the computerized records of hospitalizations, outpatient visits, and costs showed no statistically significant differences between the two groups. In the 8 months after screening, data from the hospital's computerized system showed that nonmental health outpatient visits were significantly higher in the depressed/anxious group (5.84 visits vs. 5.10; P=0.04). Our subgroup analysis of the older subjects over the median age of 42 years found those patients defined as depressed and/or anxious reported significantly more ER visits, nonmental health visits, and mental health visits during the year before the interview than those patients who were neither depressed nor anxious. The computerized records for the same time period showed no significant differences between these two older groups.
We performed a subgroup analyses stratified by gender, and the men identified as depressed and/or anxious showed a statistically significant difference in their prior year self-report of ER visits. The depressed and/or anxious women reported a statistically significant increase in the use of all services compared with the group who were not identified as depressed and/or anxious. However, for the same time period, no significant differences were identified by computerized records.
Among the 174 patients defined as depressed and/or anxious, 70 patients reported seeing a mental health provider in the past year, while 96 patients reported that they did not see a mental health provider during that year (Table 2). Those patients seeing a mental health provider reported significantly more medical visits and ER visits than depressed/anxious patients who reported that they did not see a mental health provider. The computerized records revealed that those depressed and/or anxious patients seeing mental health providers had significantly more medical outpatient visits than the patients with depression/anxiety not seeing mental health providers. The two groups had the same number of hospitalizations with slightly lower costs found in those patients seeing mental health providers, but this difference was not statistically significant. We performed a subgroup analyses stratified by age and gender and found similar patterns of utilization.
View this table:
[in this window]
[in a new window]
|
TABLE 2. Cost and utilization comparison between depressed/anxious patients with and without mental health services in the first study
|
When comparing anxious and/or depressed patients seeing mental health providers versus anxious and/or depressed patients receiving psychotropic medications not seeing mental health providers, there were no significant differences in their computerized hospital records. But through self-report, those patients seeing mental health providers had significantly more ER visits (P=0.04).
Defining depression and anxiety by the presence of current symptoms alone on the questionnaire is problematic. The absence of anxiety or depressive symptoms could reflect two different clinical states: such patients may indeed have no anxiety or depressive disorders or alternatively they may have had a past disorder that has been successfully treated. We attempted to disentangle these two clinical conditions by classifying patients as either "treatment successes" or "treatment failures." We defined "treatment successes" (n=56) as those patients without depression and/or anxiety on the MHI who reported either seeing mental health providers or taking psychotropic medication. We defined "treatment failures" (n=92) as those patients with depression and/or anxiety on the MHI who reported either seeing mental health providers or taking psychotropic medication. A third group, termed "untreated patients" (n=77), was made up of those patients with depression and/or anxiety, according to their scores on the RAND MHI, who reported not seeing mental health providers nor taking psychotropic medication (five of the patients found to be depressed and/or anxious did not answer the question about treatment history). The comparison of these three groups showed a statistically significant difference in medical utilization only on self-report, with "treatment failures" showing a higher number of nonmental health visits. We found no difference in their utilization on computer generated records.
In the second study, we examined the medical utilization of 91 BIMA patients who had been referred for psychiatric treatment. The mean age of the group was 39 years and 71% of the group were women. Twenty-two patients (24%) never came to the psychiatry clinic, 32 patients (35%) over the 1-year period had three or fewer psychiatric visits, and 37 patients (41%) had more than three visits during the next year. Although both groups of patients who had psychiatric visits had lower inpatient costs than patients with no visits, this three-group comparison was not statistically significant (Table 3). In the year after psychiatric treatment began, patients with more than three visits showed a statistically significant increase in total medical outpatient visits. We performed an analysis on a subset of older patients above the median age of 42. The overall utilization was higher for these patients but usage by psychiatric visits did not change. Gender differences were present but also did not show a difference in the pattern of utilization.
|
DISCUSSION
|
There are a number of limitations in both studies. First, we were unable to measure and control for the severity of comorbid medical or psychiatric illness. Second, we know nothing about the type of psychiatric treatment received or the adequacy of the treatment, and in our first study we do not know when, where, or for how long mental health treatment was provided. Third, in our second study we had a relatively small sample size. Fourth, in both studies the median age was relatively young, which may lessen a potential cost-offset effect. Given these limitations, our findings do confirm the difficulty in demonstrating a cost-offset effect from the use of outpatient psychiatric services, which other investigators have encountered.
For example, Klerman et al.24 and Von Korff et al.19 found that improved psychiatric treatment of depression in general medical patients did not produce an off- setting decline in the utilization of general medical services. However, the failure of such studies to demonstrate a cost-offset effect does not conclusively confirm that it does not occur, but could result from the methodological difficulties inherent in such studies. Simon and Katzelnick20 state that there is a potential for medical cost savings in treating depression but indicate that a new generation of experimental studies will be required to definitively prove cost-offset.
One difficulty inherent in naturalistic cost-offset research is that the patients receiving more psychiatric treatment are generally more severely ill psychiatrically than those receiving less care, which can result in an increased use of medical services and are therefore not totally comparable. In addition, medical care costs may not be the most suitable or appropriate outcome measure to assess the impact of psychiatric treatment on medical care. Compliance, psychological distress, or efficiency of the primary care doctor's time may be more realistic outcome measures of the effectiveness of psychiatric treatment.
In our first study, the self-report data for the patients identified as depressed and/or anxious were statistically significant for more utilization of services than those patients not depressed or anxious. This is an interesting finding from which we are unable to draw definitive conclusions, but we can entertain the possible meanings of this finding. First, the self-report data may reflect an accurate recall of the utilization in which the patients are including not only their care at BWH but also care at other medical institutions during the past year, with the latter not captured by our hospital's computer system. Second, patients' recall of visits during the past year depends on autobiographical memory that may frequently be inaccurate and biased. With event dating there is exponential decay over each unit of time that passes thus making the percent of recall after 6 months fairly low.25 When using autobiographical memory for event recall, there is a general tendency to remember events as having happened more recently than they actually did. This phenomenon, called "telescoping," leads to overreporting the frequency of events within a time period and may be a factor in the self-report data being significantly higher than the computerized data.25 Third, depressed or anxious patients' perception of their health may reflect their increased need for services, distorting their recall of service utilization.26
Both studies highlight the difficulties in identifying and treating psychiatric symptoms within a primary care setting. In the first study, less than half of those patients identified as depressed or anxious were receiving any type of psychiatric treatment. In the second study, only 41% of the patients referred to the psychiatric clinic from primary care had more than three visits. Compliance with psychiatric treatment within the primary care setting is a well- established challenge that Katon et al.27 address with a multifaceted intervention. They found that integrating mental health professionals within the primary care setting improved adherence to antidepressant medication and patient satisfaction for major depression and minor depression. These findings underscore the value of a multidisciplinary approach to the treatment of psychiatric conditions within medical settings.
Treating psychiatric illness decreases despair and psychological distress, but to expect such treatment to additionally offset medical care costs may be unrealistic. For example, we do not expect the treatment of cardiac disease to reduce overall medical costs—why then should psychiatric treatment be held to a different standard? Untreated heart disease may cause an increase in medical utilization just as untreated depression or anxiety does, but, in both cases, treatment of the underlying illness may be justified by their ability to reduce the patients' suffering and disability. The standard that psychiatric treatment must both decrease symptoms and medical costs may reflect the stigma attached to psychiatric illness, inappropriately suggesting that it should only be treated if it can be economically justified.
|
ACKNOWLEDGMENTS
|
Adapted from a presentation at the 152nd Annual Meeting of the American Psychiatric Association in Washington, D.C., May 18,1999. The authors thank Rais A. Khan, M.D. and Benjamin E. Galper for their help with data collection. Also, the authors thank Victoria Doroshenko, project analyst for CHASE Management Systems, for her help in obtaining the computer generated utilization data.
|
REFERENCES
|
-
Sartorius N, Ustun B, Silva J, et al: An international study of psychological problems in primary care. Arch Gen Psychiatry 1993; 50:819–824[Abstract]
-
Von Korff M, Shapiro S, Burke JD, et al: Anxiety and depression in primary care clinic: comparison of Diagnostic Interview Schedule, General Health Questionnaire, and practitioner assessments. Arch Gen Psychiatry 1987; 44:152–156[Abstract]
-
Katon W, Schulberg H. Epidemiology of depression in primary care. Gen Hosp Psychiatry 1992; 14:237–247[CrossRef][Medline]
-
Spitzer RL, Williams JB, Kroenke K, et al: Utility of a new procedure for diagnosing mental disorders in primary care: the PRIME-MD 1000 study. JAMA 1994; 272:1749–1756
-
Simon G, Ormel J, Von Korff M, et al: Health care costs associated with depressive anxiety disorders in primary care. Am J Psychiatry 1995; 15:352–357
-
Callahan CM, Hui SL, Nienaber NA, et al: Longtitudinal study of depression and health services use among elderly primary care patients. J Am Geriatr Soc 1994; 42:833–838[Medline]
-
Simon GE, Von Korff M, Barlow W: Health care costs of primary care patients with recognized depression. Arch Gen Psychiatry 1995; 52:850–856[Abstract]
-
Stoudemere A: Psychiatry in medical practice: implications for the education of primary care physicians in the era of managed care: part 1. Psychosomatics 1996; 37:502–508[Abstract/Free Full Text]
-
Katon W, Von Korff M, Lin E, et al: Distressed high utilizers of medical care: DSM-III-R diagnosis and treatment needs. Gen Hosp Psychiatry 1990; 12:355–362[CrossRef][Medline]
-
Duehrssen A, Jorswiek E: An empirical and statistical inquiry into the therapeutic potential of psychoanalytic treatment. Der Nervenarzt 1965; 36:166–169
-
Smith GR Jr, Monson RA, Ray DC: Psychiatric consultation in somatization disorder: a randomized, controlled study. N Engl J Med 1986; 314:1407–1413
-
Schlesinger HJ, Mumford E, Glass GV, et al: Mental health treatment and medical care utilization in a fee-for-service system: outpatient mental health treatment following the onset of a chronic disease. Am J Pub Health 1983; 73:422–429[Abstract/Free Full Text]
-
Rost K, Kashner TM, Smith RG Jr: Effectiveness of psychiatric intervention with somatization disorder patients: improved outcomes at reduced costs. Gen Hosp Psychiatry 1994; 16:381–387[CrossRef][Medline]
-
Strain JJ, Lyons JS, Hammer JS, et al: Cost offset from a psychiatric consultation-liaison intervention with elderly hip fracture patients. Am J Psychiatry 1991; 148:1044–1049
-
Strain JJ, Hammer JS, Fulop G: Academy of Psychosomatics Medicine task force on psychosocial interventions in the general hospital setting: a review of cost-offset studies. Psychosomatics 1994; 35:253–262[Abstract/Free Full Text]
-
Revicki DA, Simon GE, Chan K, et al: Depression, health-related quality of life, and medical cost outcomes of receiving recommended levels of antidepressant treatment. J Fam Pract 1998; 47:446–452[Medline]
-
Thompson D, Hylan TR, McMullen W, et al: Predictors of a medical-offset effect among patients receiving antidepressant therapy. Am J Psychiatry 1998; 155:824–827[Abstract/Free Full Text]
-
Mumford E, Schlesinger H, Glass G, et al: A new look at evidence about reduced cost of medical utilization following mental health treatment. Am J Psychiatry 1984; 141:1145–1149
-
Von Korff M, Katon W, Busch T, et al: Treatment costs, cost offset, and cost-effectiveness of collaborative management of depression. Psychosom Med 1998; 60:143–149[Abstract/Free Full Text]
-
Simon GE, Katzelnick DJ: Depression, use of medical services, and cost-offset effects. J Psychosom Res 1997; 42:333–344[CrossRef][Medline]
-
Borus JF, Olendzki MC, Kessler L, et al: The "Offset Effect" of mental health treatment on ambulatory medical care utilization and charges: month by month and grouped month analyses of a five year study. Arch Gen Psychiatry 1985; 42:573–580[Abstract]
-
Veit CT, Ware JE: The structure of psychological distress and well-being in general populations. J Consult Clin Psychol 1983; 51:730–742[CrossRef][Medline]
-
Hays RD, Stewart AL: The structure of self-reported health in chronic disease patients. J Consult Clin Psychol 1990; 2:22–30
-
Klerman GL, Budman S, Berwick D, et al: Efficacy of a brief psychosocial intervention for symptoms of stress and distress among patients in primary care. Med Care 1987; 25:1078–1088
-
Tourangeau R: Remembering what happened: memory errors and survey reports in The Science of Self-Report edited by Stone A, Turkkan J, Bachrach C, et al. Mahwah, NJ, Lawrence Erlbaum Associates, 2000, pp 29–47
-
Kihlstrom J, Eich E, Sandbrand D, et al: Emotion and memory: implications for self-report in The Science of Self-Report, edited by Stone A, Turkkan J, Bachrach C, et al. Mahwah, NJ, Lawrence Erlbaum Associates, 2000, pp 29–47
-
Katon W, Robinson P, Von Korff M, et al: A multifaceted intervention to improve treatment of depression in primary care. Arch Gen Psychiatry 1996; 53: 924–932
This article has been cited by other articles:
|
|
|
|
 
G. Schneider, M. Wachter, G. Driesch, A. Kruse, H.-G. Nehen, and G. Heuft
Subjective Body Complaints as an Indicator of Somatization in Elderly Patients
Psychosomatics,
April 1, 2003;
44(2):
91 - 99.
[Abstract]
[Full Text]
[PDF]
|
|
|
Get information about faster international access.
Privacy Policy
Copyright © 2000
Academy of Psychosomatic Medicine.
All rights reserved.
Home
| Search
| Current Issue
| Past Issues
| Subscribe
| All APPI Journals
| Help
| Contact Us
|
TOP
ABSTRACT
INTRODUCTION
