Psychosomatics 39:431-436, October 1998
© 1998 The Academy of Psychiatric Medicine
Clinical Predictors of Mental Disorders Among Medical Outpatients
Validation of the "S4" Model
Jeffrey L. Jackson, M.D., M.P.H.,
Patrick G. O'Malley, M.D., and
Kurt Kroenke, M.D.
Received December 7, 1997; revised February 3, 1998; accepted February 11, 1998. From the Departments of Medicine, Uniformed Services University of the Health Sciences (USUHS), Bethesda, MD; the Walter Reed Army Medical Center, Washington, DC; and the Regenstrief Institute for Health Care and Indiana University School of Medicine, Indianapolis, IN. Address reprint requests to Dr. Jackson, Dept of Medicine–EDP, USUHS, 4301 Jones Bridge Rd., Bethesda, MD 20814.
|
ABSTRACT
|
The authors previously reported four clinical cues that predicted a subgroup of ambulatory patients likely to have depressive and anxiety disorders. The authors' purpose in this study was to validate this model in another cohort of 185 consecutive adult referrals to a rheumatology clinic. The authors found 4 variables important in predicting mental disorders: recent stress (odds ratio [OR]: 3.3, 95% confidence interval [CI]: 1.5–7.1); >5 somatic symptoms (OR: 4.5, 95% CI: 1.1–9.5); only fair or poor health status (OR: 3.4, 95% CI: 1.6–7.4); and symptom severity (OR: 1.6, 95% CI: 0.8–3.6). There was a stepwise increase in the likelihood of a mental disorder with an increasing number of predictors. The authors conclude that these clinical cues may allow clinicians to select patients in which formal screening for mental disorders would be particularly fruitful.
Key Words: clinical predictors • mental disorders • S4 model
|
INTRODUCTION
|
Mental disorders are present in up to a third of patients seen in the outpatient setting,1–4 yet despite the availability of a number of case-finding instruments,2,5,6 such disorders frequently remain undiagnosed.2,3,7 Screening all outpatients for mental disorders has not been shown to be cost-effective or feasible. An alternative strategy would be to target case-finding patients at higher risk, especially given the time constraints and competing demands of primary care clinicians.8,9
In a prior study of 500 patients presenting to an ambulatory clinic with physical complaints, we discerned four clinical predictors—recent stress, severity of the presenting symptom, total number of physical symptoms, and poor self-rated health status—that significantly increased the likelihood of a depressive or anxiety disorder.4 These four factors, which we identified as the "S4" model (Stress, Severity, Symptom count, self-rated health Status), often became apparent during the normal course of an interview and identify a population of patients in whom psychiatric screening may be particularly warranted.
In this study, conducted in 1997, we sought to validate this predictive model in a cohort of patients referred to a rheumatology clinic. This study population was chosen for two reasons. First, a musculoskeletal complaint was the most common symptom in our prior study of patients presenting to a primary care walk-in clinic (36%). Second, we were interested in determining how well this model would generalize to a subspecialty population.
|
METHODS
|
Consecutive adult referrals to the rheumatology clinics at either the Walter Reed or Madigan Army Medical Centers were invited to participate. These rheumatology clinics serve large, geographically defined populations and receive referrals from numerous outlying clinics as well as from the medical centers' ambulatory care clinics. These ambulatory clinics provide primary care for both active-duty and retired military personnel and their families. Physicians are free to refer patients with rheumatologic complaints when the doctors deem further evaluation is warranted. The case mix as well as age and gender distribution of patients cared for in a U.S. Army medical center is similar to that seen in civilian internal medicine practices.10 The protocol was approved by the Walter Reed and Madigan Clinical Investigation Committees, and all study participants provided signed informed consent.
Before seeing the rheumatologist, each patient completed a questionnaire on symptom characteristics; symptom-related expectations of care; and self-rated health status (excellent, very good, good, fair, or poor). Psychiatric disorders were established by using a self-administered version of PRIME-MD. PRIME-MD is a validated instrument that yields DSM-IV criteria-based diagnoses.2 The overall diagnostic accuracy and interobserver agreement of PRIME-MD are comparable to other structured psychiatric interviews administered by mental health specialists.11 With the self-administered version of PRIME-MD, depressive and anxiety disorders are diagnosed, according to standard DSM-IV criteria.
Statistical analyses were done by using Stata.12 Categorical variables were analyzed by using chi-square, and continuous variables were assessed with Students' t-test or Kruskal-Wallis testing, as appropriate. Our primary outcome variable was the presence of a mental disorder. Variables significant on univariate screen at P<0.20 were identified as potential predictive variables and fit into a logistic-regression model after the methods of Hosmer and Lemeshow.13 The performance of individual predictors and the overall model was compared with the model's original performance in the derivation set of 500 general medical patients.4
|
RESULTS
|
Patient Demographics
The 185 patients had a mean age of 46 (range: 18–87); 64% were women; 65% were white, 19%, African American, 6%, Hispanic, and 5% Asian; and 29% were college graduates. The median symptom duration was 730 days. Eighty-four percent (n=155) of the patients were given a diagnosis on the initial visit by the rheumatologist. Fifty-seven (31%) had connective tissue diseases, such as rheumatoid arthritis, Sjögren's syndrome, or systemic lupus erythematosus. Sixty-three (34%) had articular processes, such as gout or degenerative joint disease, and 35 (19%) were diagnosed with nonarticular, noninflammatory processes, such as fibromyalgia or myofascial pain syndrome. The symptom had been present less than 1 month in only 2% of the patients, and only 20% reported duration less than 6 months. Many patients were worried their symptoms could indicate an underlying serious illness (67%) or interfere substantially with their usual activities (52%). Most patients identified one or more previsit expectations: 79% wanted an explanation of what was causing their symptoms; 59% an estimate of how long their symptoms would last; 46% a prescription; 36% a diagnostic test; 12% another referral; and 19% something else (e.g., reassurance or a treatment program).
Prevalence of Psychiatric Disorders
Sixty-one patients (34%) had a threshold or subthreshold depressive or anxiety disorder (Table 1), and the prevalence of each was similar between the two rheumatology clinics. Forty-eight patients (27%) had an anxiety disorder, with 7 (4%) having panic disorder. Forty-six (25%) patients had a depressive disorder, with 13 (7%) meeting criteria for major depression. Thirty-three (18%) patients had both a depressive and anxiety disorder.
There were no differences between the patients with and without psychiatric disorders in demographics (age, gender, ethnicity, marital status, educational attainment), symptom duration, or patient-reported limitation in activities as a result of symptoms (Table 1). There were also no differences in the number or type of previsit expectations. However, patients with psychiatric disorders were more likely to report serious illness worry (odds ratio [OR]: 2.2, 95% confidence interval [CI]: 1.0–4.6); recent stress (OR: 3.6, 95% CI: 1.8–7.1); higher symptom severity (6.7 vs. 5.9, P<0.0001); more somatic symptoms (7.9 vs. 4.4, P<0.0001); and worse overall health status (P<0.0001) (Table 1).
Regression Modeling
Variables that distinguished the patients with and without mental disorders on univariate screen at P<0.20 included age, gender, symptom duration and severity, recent stress, number of somatic symptoms, serious illness worry, and overall health status. Patient reports of recent stress and serious illness worry were found to be unacceptably collinear (R=0.47), and the better-fitting variable (stress) was retained. The final model is shown in Table 2. Variables independently significant in predicting the presence of an underlying mental disorder included recent stress (OR: 3.3, 95% CI: 1.5–7.1), more than 5 somatic symptoms (OR: 4.5, 95% CI: 1.1–9.5), and self-reported health as only "fair" or "poor" (OR: 3.4, 95% CI: 1.6–7.4). Symptom severity was retained in the model because of its statistically significant contribution to the predictive model but was not independently significant (OR: 1.6, 95% CI: 0.8–3.6). There was a stepwise increase in the likelihood of a mental disorder with increasing number of predictors: patients with 0, 1, 2, 3, or 4 predictors had a likelihood of a depressive or anxiety disorder of 0%, 16%, 37%, 70%, and 94%, respectively (Table 3). The receiver operator curve (ROC) area was 0.88, identical to the ROC area in the original derivation sample of 500 general medical patients.4 By using regression coefficients, 78% of the patients were correctly classified as either having or not having a mental disorder. Finally, sensitivity and specificity were calculated for different thresholds of the four-predictor "S" model. A threshold of  1, 2, 3, and 4 predictors had a sensitivity for depressive or anxiety disorders of 100%, 89%, 63%, and 21%, respectively, and a specificity of 15%, 55%, 86%, and 99%, respectively. For example, 63% of the patients with a depressive or anxiety disorder diagnosed by PRIME-MD had three or more clinical predictors, whereas 89% of the patients without a disorder had less than three predictors.
|
DISCUSSION
|
Depressive or anxiety disorders were present in one-third of new referrals to rheumatology clinics. The "S4" model (Symptom count, Stress, Severity, health Status) predicted patients with underlying depressive and anxiety disorders, validating the results of our prior study. There was a stepwise increase in the percentage of patients with mental disorders, from none among the patients with no predictor present, to 94% among the patients with all four predictors. Despite the shift from a primary care setting to a subspecialty clinic, the remarkable similarity of specific predictive variables and their respective ORs suggests robustness and, hence, generalizability of the model (Table 2).
The exact questions in our study were 1) "During the past week have you been under stress?", scored as yes or no, 2) "Describe how bad your symptom is, from 10 (unbearable) to 0 (none at all)," scored as positive for responses greater than 5; and 3) "In general, would you say your health is excellent, very good, good, fair, or poor?", scored as positive for fair or poor responses.4 Physical symptom count was measured with the PRIME-MD 15-symptom checklist, with endorsement of 6 or more symptoms considered positive for this model. During clinic encounters, multiple somatic complaints often become apparent early in the interview and might serve as a surrogate for a formal symptom checklist.
A number of barriers to recognizing mental disorders in the busy outpatient setting have been identified, including time, stigmatization, and somatization. The majority of primary care visits are completed in less than 15 minutes.14 Given the limited time, coupled with an expanded agenda of tasks expected from primary care physicians, case finding in selected patients may be more feasible than screening every clinic attendee.
The "S4" model is not suggested as a substitute for direct inquiry about items that permit criteria-based diagnosis of either a depressive, anxiety, or somatoform disorder. Preliminary evidence suggests that simply asking a few questions about depressed mood or anhedonia is quite sensitive (over 90%) in screening for depressive disorders,15 although single-item screens for anxiety disorders have not been studied. While not replacing diagnostic questions, S4 items may nonetheless be clinically useful for several reasons. First, if one or more of the items spontaneously surface during the clinical interview of a medical patient, they may serve as "red flags" for a potential underlying mental disorder. Recent stress, multiple physical complaints, symptom severity disproportionate to clinical findings, or low ratings of overall health identify patients at higher risk of underlying psychopathology in whom specific probing about mental disorders is particularly warranted.
Second, most patients who present with mental disorders in the medical setting report physical rather than emotional symptoms,4,16–20 and premature questioning about depression or anxiety can elicit negative reactions in some patients (e.g, "So you are saying this is all in my head?"). In such patients, a gradual approach to linking physical complaints and psychological symptoms may be more effective. Indeed, the causal relationship (and directionality) between physical and emotional symptoms is not clear-cut in many medical patients. Therefore, the physician need not always insist on depression or anxiety as the "cause" of the patient's physical symptoms, but rather can adopt a more neutral position by considering the depressive or anxiety symptoms as simply a coexisting condition, or even a potential consequence of persistent physical distress. For example, to the patient with back pain who is also positive for some of the S4 items, the physician might say, "You told me that your pain is particularly severe. With pain this bad, many patients also can feel a bit down or depressed. Has this happened to you?" Or "Besides the back pain, you mentioned that the past month has been particularly stressful. With all of this going on, have there been times when you felt anxious or worried or a bit down?"
Our study has several limitations. The initial sample was a group of patients presenting to a primary care walk-in clinic with a physical complaint, whereas the validation sample was a group of patients referred to a subspecialty clinic. While the model behaved remarkably similarly in those settings, it would be helpful to test it in a sample of patients within the context of continuity of care visits. It should be noted, however, that previous studies have shown a prevalence of mental disorders similar to that found in both of our cohorts.11,21–25 Second, a substantial proportion of patients had subthreshold diagnoses, such as minor depressive and anxiety disorders, for which antidepressants or other treatment have been inadequately studied. While it is likely the natural history for these disorders is more favorable, it has been found that such patients have higher rates of unmet expectations, are more likely to be reported as "difficult" by their provider, and have psychological symptoms that persist for months after the initial encounter.4 Therefore, identification of such patients may be important.
The number of predictors patients had from the S4 model (Symptom count, Stress, Severity, health Status) stratified them into groups ranging from a very low to a very high likelihood of a depressive or anxiety disorder. Since many busy clinicians may not consider it either feasible or desirable to screen all patients, the presence of one or more of the S4 clinical clues may be one marker that could identify selected patients for psychiatric case finding.
|
ACKNOWLEDGMENTS
|
The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting those of the Department of the Army or the Department of Defense.
|
REFERENCES
|
-
Ormel J, Von Korff M, Uston TB, et al: Common mental disorders and disability across cultures: results from the WHO Collaborative Study on Psychological Problems in General Health Care. JAMA 1994; 272:1741–1748[Abstract]
-
Spitzer RL, Williams JBW, 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[Abstract]
-
Sartorius N, Goldberg D, de Girolamo G, et al (eds): Psychological Disorders in General Medical Settings. Lewiston, NY, Hofgrebe & Huber, 1990
-
Kroenke K, Jackson JL, Chamberlin J: Depressive and anxiety disorders in patients presenting with physical complaints. Am J Med 1997; 103:339–347[Medline]
-
Mulrow CD, Williams JW, Gerety MB, et al: Case-finding instruments for depression in primary care settings. Ann Intern Med 1995; 122:913–921[Abstract/Free Full Text]
-
Broadhead WE, Leon AC, Weissman MM, et al: Development and validation of the SDSS-PC screen for multiple mental disorders in primary care. Arch Fam Med 1995; 4:211–219[Abstract]
-
Ormel J, Koeter MWJ, van den Brink W, et al: Recognition, management, and course of anxiety and depression in general practice. Arch Gen Pychiatry 1991; 48:700–706.
-
Klinkman MS: Competing demands in psychosocial care: a model for the identification and treatment of depressive disorders in primary care. Gen Hosp Psychiatry 1997; 19:98–111[Medline]
-
Kroenke K: Discovering depression in medical patients: reasonable expectations. Ann Intern Med 1997; 126:463–465[Free Full Text]
-
Johnson JE, Jenkins TR, Pinholt EM, et al: Content of ambulatory internal medicine practice in an academic army medical center and an army community hospital. Milit Med 1988; 153:21–25
-
Spitzer RL, Williams JBW, Kroenke K, et al: The PRIME-MD 1000 study: description, validation, and clinical utility of a new procedure for diagnosing mental disorders in primary care. JAMA 1994; 272:1749–1756
-
StataCorp: 1995 Stata Statistical Software: Release 5.0. College Station, TX, Stata Corporation, 1995
-
Hosmer DW, Lemeshow S. A goodness-of-fit test for the multiple logistic regression model. Communications in Statistics 1980; 10:1043–1069
-
Schappert SM: National Ambulatory Medical Care Survey: 1989 summary. National Center for Health Statistics. Vital Health Stat 1992; 13:1–17
-
Whooley MA, Avins AL, Miranda J, et al: Case-finding instruments for depression: two questions are as good as many. J Gen Intern Med 1997; 12:439–445[Medline]
-
Katon W, Kleinman A, Rosen G: Depression and somatization: a review. Am J Med 1982; 72:127–135, 241–247[Medline]
-
Bridges KW, Goldberg DP: Somatic presentations of DSM-III psychiatric disorders in primary care. J Psychosom Res 1985; 29:563–569[Medline]
-
Kirmayer LJ, Robbins JM: Three forms of somatization in primary care: prevalence, co-occurrence and sociodemographic characteristics. J Nerv Ment Dis 1991; 179:647–655[Medline]
-
Lipowski ZJ: Somatization and depression. Psychosomatics 1990; 31:13–21[Abstract/Free Full Text]
-
Marple RL, Kroenke K, Lucey CR, et al: Concerns and expectations in patients presenting with physical complaints: frequency, physician perceptions and actions, and two-week outcome. Arch Intern Med 1997; 157:1482–1488[Abstract]
-
Kroenke K, Spitzer RL, Williams JB, et al: Physical symptoms in primary care. Predictors of psychiatric disorders and functional impairment. Arch Fam Med 1994; 3:774–779[Abstract]
-
Barrett JE, Barrett JA, Oxman TE, et al: The prevalence of psychiatric disorders in primary care. Arch Gen Psychiatry 1988; 45:1100–1106[Abstract]
-
Schulberg HC, Burns BJ: Mental disorders in primary care: epidemiologic, diagnostic and treatment direction. Gen Hosp Psychiatry 1988; 10:79–87[Medline]
-
Kroenke K, Price R: Symptoms in the community. Prevalence, classification and psychiatric comorbidity. Arch Intern Med 1993; 153:2474–2480[Abstract]
-
Philbrick JT, Connelly JE, Wofford AB: The prevalence of mental disorders in rural office practice. J Gen Intern Med 1996; 11:9–15[Medline]
This article has been cited by other articles:
|
|
|
|
 
J. L. Jackson, J. Chamberlin, and K. Kroenke
Gender and Symptoms in Primary Care Practices
Psychosomatics,
October 1, 2003;
44(5):
359 - 366.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
B. Lowe, K. Grafe, K. Kroenke, S. Zipfel, A. Quenter, B. Wild, C. Fiehn, and W. Herzog
Predictors of Psychiatric Comorbidity in Medical Outpatients
Psychosom Med,
September 1, 2003;
65(5):
764 - 770.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. L. Jackson, J. S. Houston, S. R. Hanling, K. A. Terhaar, and J. S. Yun
Clinical Predictors of Mental Disorders Among Medical Outpatients
Arch Intern Med,
March 26, 2001;
161(6):
875 - 879.
[Abstract]
[Full Text]
[PDF]
|
|
|
Get information about faster international access.
Privacy Policy
Copyright © 1998
Academy of Psychosomatic Medicine.
All rights reserved.
Home
| Search
| Current Issue
| Past Issues
| Subscribe
| All APPI Journals
| Help
| Contact Us
|
TOP
ABSTRACT
INTRODUCTION
