Letter

"Trumping Rules" Affect Diagnoses of Persons With Severe Mental Illness

To the Editor: Accurately assessing comorbid psychiatric conditions among patients with severe mental illness is important in light of data suggesting that secondary disorders, such as anxiety disorders, can aggravate the clinical course of severe mental illnesses ( 1 ). Nevertheless, some studies suggest that comorbid conditions are often overlooked or inaccurately diagnosed in community mental health clinics ( 2 ).

We assessed the frequency of anxiety disorders (panic disorder, agoraphobia, social phobia, generalized anxiety disorder, and posttraumatic stress disorder [PTSD]) in two samples of patients at a state-funded community mental health center. Sample 1 consisted of 27 patients with severe mental illness attending a day hospital who were randomly selected from a program roster and interviewed between January 2002 and January 2004. Sample 2 consisted of 24 outpatients with severe mental illness and probable PTSD (that is, treatment seekers) who were interviewed between March 2004 and December 2005. We examined concordance rates between agency records (chart data spanning the length of each patient's care) and the Mini-International Neuropsychiatric Interview (MINI), a structured psychiatric interview ( 3 ). The study was conducted with full approval from relevant institutional review boards, and written informed consent was obtained for all participants.

Patients from both samples met criteria for severe mental illness—that is, they had a diagnosis of schizophrenia, bipolar disorder, or major depression and required assistance with independent living. The mean±SD age of sample 1 was 46.74±13.53; 21 of the 27 patients (78%) were African American, and six (22%) were Caucasian. Sixteen patients (59%) in sample 1 were male, and two (7%) were married. In sample 2 the mean age was 41.92±7.87; ten of the 24 patients (42%) were African American, 13 (54%) were Caucasian, and one (4%) was biracial. Seven patients (29%) in sample 2 were male, and one (4%) was married. No patient in either sample worked full-time.

Although a number of anxiety disorders were identified in structured interviews, none were recorded in patients' charts. In both samples, some patients had more than one anxiety disorder. In sample 1 six patients (22%) were diagnosed as having an anxiety disorder, three (11%) had PTSD, two (7%) had agoraphobia, two (7%) had generalized anxiety disorder, one (4%) had social phobia, and one (4%) had panic disorder. Thirteen patients in sample 2 (54%) had an anxiety disorder; PTSD was not included because the sample was selected for probable PTSD. Nine patients (38%) had panic disorder, six (25%) had agoraphobia, six (25%) had social phobia, and six (25%) had generalized anxiety disorder.

Results indicate that primary diagnoses of severe mental illness may overshadow the detection of other conditions, such as anxiety disorders. This may stem from the hierarchical classification of psychosis and neurosis, in which disorders higher in the hierarchy of organicity "trump" lower-order disorders ( 4 ).

Some limitations of the study should be noted. First, it is possible that clinicians at the sample sites were aware of and treating these anxiety disorders and simply did not document them. In addition, rates of PTSD in sample 1 were low compared with rates in other studies ( 5 ), which raises questions about the sensitivity of the MINI compared with other measures of PTSD.

Despite these limitations, accurate identification and documentation of secondary diagnoses are critical to ensuring that patients receive fully integrated care. Accurate chart documentation is particularly important in public-sector settings with high clinician turnover, because poor documentation can affect continuity of care. More rigorous screening and documentation practices may be needed in these settings to ensure that patients receive services for secondary disorders that have distressing symptoms or that exacerbate symptoms of severe mental illness.

Acknowledgments and disclosures

This work was partially supported by grants MH-074468, MH-065248, and MH-65517 from the National Institute of Mental Health. The authors gratefully acknowledge the contributions of David Sheil, M.S.W., and Deborah Dinovo, M.Ed., L.P.C.S. The authors report no competing interests.