Since the revision of DSM-III, high rates of co-occurring psychiatric disorders have been observed, particularly in cases of moderate and severe psychiatric illness. The reason lies in the design of the diagnostic system itself: DSM-IV is a descriptive, categorical system that splits psychiatric behaviors and symptoms into numerous distinct disorders, and uses few exclusionary hierarchies to eliminate multiple diagnoses.
Since the revision of DSM-III, high rates of co-occurring psychiatric disorders have been observed, particularly in cases of moderate and severe psychiatric illness.1 The reason lies in the design of the diagnostic system itself: DSM-IV is a descriptive, categorical system that splits psychiatric behaviors and symptoms into numerous distinct disorders, and uses few exclusionary hierarchies to eliminate multiple diagnoses. In this article, we examine the evolution of our current diagnostic system to better understand this emergence of comorbid psychiatric diagnoses. Using clinical examples, we explore strategies that can be implemented in revisions of the DSM for reducing psychiatric comorbidity, and we consider the strengths and weaknesses of each.
PSYCHIATRIC LIMITATIONS OF COMORBIDITY
Comorbidity is often used to describe the presence of more than one identified psychiatric disorder in a patient. Feinstein2 first coined the term "comorbidity" in the general medical literature, defining it as "any distinct additional clinical entity that has existed or that may occur during the clinical course of a patient who has the index disease under study" (eg, a patient with asthma and diabetes mellitus).
But in psychiatry, can we distinguish clinical entities in this way? Is the person who binges and purges and also drinks to excess suffering from 2 distinct disorders (bulimia nervosa and alcohol abuse) or a single disorder of impulse control? Similarly, when symptoms of generalized anxiety and depression co-occur, whether they indicate the presence of 2 distinct clinical entities or are 2 components of a single disorder is mostly a matter of speculation. Thus, the term "comorbidity" is misleading in psychiatry, because it implies that we have identified multiple distinct disease states.
Recognizing the limits of our understanding, DSM-IV guides clinicians to convey the maximum amount of descriptive information possible, even if a diagnosis of numerous simultaneous disorders results. According to DSM-IV, a patient who meets criteria for dysthymia, abuses alcohol, has obsessions and compulsions, and binges and purges would be assigned 4 separate Axis I disorders. Assigning multiple diagnoses allows for the communication of clinical complexity that a more general, all-encompassing diagnosis (such as "severe neurotic disorder") lacks.
ENCOURAGING COMORBID DIAGNOSES
While the DSM-IV diagnostic scheme was designed to capture clinical complexity via multiple diagnoses, there is evidence that this diagnostic complexity is not being adequately captured in medical records systems and everyday clinical practice. Zimmerman and Mattia3 have reported that clinicians routinely underdetect psychiatric comorbidity compared with research assessments using structured diagnostic interviews. One study found that 5 times as many comorbid diagnoses were made when using semistructured interviews compared with clinicians' assessments alone.4 The reasons underlying this discrepancy are most likely complex. With shrinking reimbursement and rapid patient flow, psychiatrists may lack sufficient time to undertake complete diagnostic assessments. For practical purposes, clinicians may also think more reductively when approaching patients, dismissing diagnoses that seem unrelated to the primary problem that resulted in the patient seeking care.
What are the costs of a system that encourages the concept of comorbidity? Recording 5 or 6 diagnoses on a patient's chart may obscure the intended focus of treatment, particularly for nonpsychiatric providers who request a consultation for their patient. It may confuse or alienate patients to receive diagnoses of multiple disorders. Furthermore, it may overwhelm medical records systems or administrative databases. Many health information systems only allow for coding a limited number of diagnoses, with the result that any additional diagnoses are ignored. For example, 3 patients with a primary diagnosis of major depressive disorder may have very different comorbid diagnoses (eg, obsessive-compulsive disorder, alcohol dependence, or posttraumatic stress disorder). A system that records only the major depression would imply that these 3 patients were diagnostically homogeneous, when in fact their co-occurring diagnoses suggest otherwise.
Accounting for these additional diagnoses often provides rich diagnostic information that is relevant in designing individualized treatment plans. For example, the patient with depression and an additional diagnosis of panic disorder may benefit from a short course of a benzodiazepine until an SSRI has taken therapeutic effect. Conversely, a clinician might be expected to hesitate before taking the same step in a depressed patient with a co-occurring substance abuse disorder.
At its best, our current diagnostic system has the potential to communicate large amounts of clinical information about patients with complex problems, allowing for targeted treatments and precisely defined study populations. At its worst, it can be overwhelming to clinicians and health information systems, elicit skepticism from patients and clinicians in other fields who do not understand the system, and obscure the focus of our treatments by "losing the forest for the trees."
DSM AND CO-OCCURRING DIAGNOSES
The original version of DSM was a descriptive system that incorporated many of the concepts and the structure of Emil Kraepelin's classifications of mental disorders. Compared with subsequent revisions, DSM-I and DSM-II followed a "one disease-one diagnosis" model. The clinician strove to diagnose parsimoniously, using qualifying phrases (such as "with neurotic reaction" and "with psychotic reaction") to describe complex cases. Implicit in these early DSM systems were many assumptions about the etiology and nature of mental disorders.
DSM-III, however, took a different approach of splitting diagnoses into a large number of relatively narrowly defined psychiatric disorders, providing operationalized criteria for each. For example, DSM-III split the single DSM-II category "phobic neurosis" into 5 distinct DSM-III categories. DSM-III did not encourage practitioners to assume that a wide variety of phobic reactions (such as separation anxiety disorder and simple phobia) were simply variants of the same disorder. Not surprisingly, from DSM-II to DSM-IV, the number of distinct psychiatric diagnoses nearly doubled.
Diagnostic splitting comes at the risk of spurious comorbidity (attributing a single set of symptoms toward criteria for several disorders). Symptoms such as sleep disturbance, psychomotor changes, and poor concentration in the setting of depression can also, theoretically, be counted as symptom criteria for other conditions described in DSM,increasing the likelihood of multiple diagnoses. One strategy used in DSM-III to reduce spurious comorbidity arising from diagnostic splitting was the introduction of diagnostic hierarchies. For example, the criteria for agoraphobia indicate that the diagnosis should not be given if the characteristic avoidant behavior is really due to obsessive-compulsive disorder.
However, the use of the phrase "due to" forces the clinician to determine when a symptom is attributable to one disorder versus another. As mentioned earlier, currently such decisions are based on assumptions about causality that are not empirically based. Resulting partly from research conducted by Boyd and colleagues5 in the 1970s and 1980s, many assumptions about the relationship between mood and anxiety symptoms came into question. In subsequent revisions of DSM, starting with DSM-III-R, these concerns led to the removal of many exclusionary hierarchies.
The ultimate result of this combination of widespread diagnostic splitting, with few exclusionary hierarchies, is that patients qualify for multiple diagnoses. In a study involving 500 patients presenting for intake in a general psychiatric clinic, Zimmerman and Mattia,6 using semistructured clinical interviews, noted that more than a third of the patients qualified as having 3 or more Axis I disorders.
STRATEGIES TO REDUCE COMORBID DIAGNOSES
A prominent example of excessive comorbidity in DSM-IV that has resulted in dissatisfaction among clinicians is in the area of personality disorders.7,8
DSM describes 10 personality disorders, with no hierarchical system provided to reduce comorbidity. According to this system, when an individual meets criteria for more than 1 disorder, all diagnoses should be listed. Widiger and colleagues8 have pointed out that certain psychiatric inpatients meet criteria for 3 to 5, and in some cases 7, personality disorders. Using pooled data for 1116 inpatients and outpatients, Stuart and colleagues9 found that those patients who met criteria for any personality disorder actually met criteria for an average of 2.7 personality disorders.
However, clinicians still tend to diagnose personality disorders reductively. In a study in which psychiatrists were given case histories of 46 patients who met criteria for 4 personality disorders, two thirds diagnosed only 1 personality disorder, a quarter diagnosed 2, and none diagnosed all 4.10
An alternative approach to the DSM-IV categorical method for diagnosing personality disorders is to adopt a dimensional model that depicts relative psychopathology as points on a spectrum. The 5-factor model (FFM) of personality is one externally validated dimensional system that has been proposed as an alternative to the DSM-IV categorical system.11,12 Rather than applying distinct criteria to distinguish "case" from "noncase," the FFM is used to rate patients on 5 dimensions: neuroticism, extraversion, openness to experience, agreeableness, and conscientiousness. By using the dimensional model, the FFM is able to characterize patients in complex ways without applying multiple labels that imply distinct disorders.
For example, a patient with psychopathology suggestive of borderline, antisocial, and narcissistic personality disorders in the DSM-IV system would instead be characterized as high in neuroticism (with corresponding subscales reflecting propensity for anger, irritability, stress tolerance, and so forth), low in agreeableness (antagonistic), and high in openness to experience (exaggerated mood states, preoccupation with fantasy).
However, the lack of discrete categories does not lend itself to the study of (ostensibly) distinct clinical populations, and also does not provide a straightforward answer to the question, "does this patient have a personality disorder?" Providing such a categorical answer to this question is important both for treatment planning (which often requires a categorical judgment of whether to treat the patient), and for practical concerns such as determining eligibility for disability. Moreover, most health information systems (eg, for clinical information or insurance) are not equipped to incorporate dimensional approaches.
Another diagnostic strategy that would reduce comorbidity is to impose additional diagnostic hierarchies. As mentioned above, exclusionary hierarchies are based on the concept that 1 diagnosis takes precedence over 1 or more subordinate diagnoses. The underlying assumption is that the symptoms of the subordinate diagnosis are associated features of the primary disorder (and thus do not warrant an additional psychiatric diagnosis). Some have recommended that certain Axis II diagnoses, such as borderline personality disorder, take precedence over others, such as dependent or histrionic personality disorder.13
The problem with imposing diagnostic hierarchies is that they imply a certain knowledge of symptom attribution that is at odds with the descriptive approach of DSM. Some of the remaining hierarchies, such as the exclusion of generalized anxiety disorder in the setting of comorbid major depressive disorder, have continued to draw criticism.14 While expanding the number of diagnostic hierarchies in DSM would certainly reduce comorbidity, this assumes an understanding of the etiologies of mental illnesses and increases the risk of losing clinically relevant distinctions in complex cases.
Mixed and combined diagnostic categories
Another strategy that has been used to address comorbidity is the development of mixed diagnostic categories, which lump together categories that have been separated in DSM (eg, combining major depressive disorder and generalized anxiety disorder into the single category, mixed anxiety/depression).15 One drawback of combining single categories into combined categories based on known co-occurring syndromes is that the number of possible combinations and permutations of categories (in a mathematical sense at least) could easily reach into the thousands.
Another way to reduce diagnostic comorbidity is to combine separate categories into higher-order constructs. One example of a "lumping" diagnostic strategy already in use in DSM-IV is combining the 10 specific personality disorders of DSM into 3 personality disorder clusters based on presumed common characteristics. Clinicians commonly incorporate these clusters into diagnostic formulations, such as personality disorder not otherwise specified, with cluster B traits; researchers have used these mixed categories to delineate patient populations in studies of Axis I comorbidity, treatment responsiveness, and prognosis.16
The co-occurrence of personality disorders within each cluster is generally higher than between-cluster comorbidity. For example, in the previously mentioned study by Stuart and colleagues,9 73% of patients in whom narcissistic personality disorder was diagnosed also met criteria for histrionic personality disorder, and many met criteria for a third and fourth diagnosis. Under the cluster system, comorbidity of personality disorders is reduced because the clinician would simply note "cluster B personality disorder" regardless of the number of specific disorders actually present. However, currently there is insufficient research to justify lumping the personality disorders into clusters for all diagnostic purposes.
In a seminal article on psychiatric diagnosis, Robins and Guze17 posited that validity could be enhanced via more precise clinical description, delineation of syndromes, treatment response profiles, and biological correlates. This was based on the expectation that empirical evidence would eventually become the mainstay of psychiatric diagnosis. More than 3 decades later, and just a few years before the publication of DSM-V, our understanding of the etiology and pathogenesis of mental illness is still very limited. The research that will fully illuminate our understanding of mental illness is still many years down the road.
As a field, we in psychiatry need to address the implications of our lack of understanding about pathophysiology for the use of our current diagnostic system. The "atheoretical" approach of DSM explicitly acknowledges the limits of our understanding and encourages co-occurring diagnoses in the hope that the maximum amount of clinically relevant information will be captured. Unfortunately, this also makes the current system quite cumbersome to use as it was intended. Clinicians and health information systems intrinsically place certain information at a higher order of importance, and they use clinical judgment to prioritize and perhaps omit diagnoses. Consequently, clinicians may fail to account for or communicate diagnostic complexity where it exists. Because of the unwieldy nature of making numerous diagnoses, the application of the DSM system in the real world may be veering toward more idiosyncratic use. More research is needed if psychiatry, as a field, aspires to accuracy and uniformity in diagnosis. The ultimate goal is to increase the clinical usefulness of DSM to provide better case conceptualization, communication, and accuracy of prognosis.18
Given that the next edition of DSM will be saddled with many of the same limitations as its predecessors, some practical modifications may lessen the burden of multiple diagnoses. Accepting that the maximal comorbidity strategy is too cumbersome, we could consider modifying the current multiaxial system. Rather than directing providers to list every Axis I diagnosis for which the patient meets criteria, providers could report the 1 or 2 that are the immediate focus of care, followed by a brief list of target symptoms and relevant clinical issues.
For example, a primary diagnosis of major depressive disorder could be followed by a target symptom list of "depressed mood and insomnia (severe); panic attacks (moderate); homelessness (in shelter)." In such a system, defining whether a third or fourth set of symptoms meets independent diagnostic criteria (such as whether the panic attacks are a facet of the major depressive disorder or their own clinical entity) is less important than efficiently communicating and recording diagnostic and nondiagnostic information. This may, in fact, more closely resemble how pro-viders think about treatment: starting with diagnosis, but tracking individual symptoms to monitor the effectiveness of interventions. Such a model would defer to the providers what constitutes clinically relevant information. Again, how to guide providers in making such differentiations in a consistent manner is not immediately clear and requires future investigation.
In summary, the current strategy of diagnosing maximal comorbidity may not be optimal. The practice of listing multiple diagnoses has the power to both enhance and obscure important clinical information. As DSM-V is developed, opportunities for reducing comorbidity by lumping diagnoses (eg, replacing the 8 specific paraphilias with a single disorder); formalizing conventions for omitting nonrelevant diagnoses; using a dimensional system to characterize personality pathology; and simplifying or eliminating the multiaxial system and using one that resembles a list of the kind more commonly used in medical practice should be explored. As we move forward in addressing these challenges, we must strive to implement explicit decision rules based on the best data available or risk reverting to subjective and impressionistic formulations like those used more than half a century ago in DSM-I.
Kessler RC, Chui WT, Demler O, Walters EE. Prevalence, severity, and comorbidity of 12-month
disorders in the National Comorbidity Survey Replication.
Arch Gen Psychiatry
Feinstein AR. The pre-therapeutic classification of co-morbidity in chronic disease.
J Chronic Dis.
Zimmerman M, Mattia JI. Principal and additional
disorders for which outpatients seek treatment.
Basco MR, Bostic JQ, Davies D, et al. Methods to improve diagnostic accuracy in a community mental health setting.
Am J Psychiatry.
Boyd JH, Burke JD Jr, Gruenberg E, et al. Exclusion criteria of
a study of co-occurrence of hierarchy-free syndromes.
Arch Gen Psychiatry.
Zimmerman M, Mattia JI. Psychiatric diagnosis in clinical practice: is comorbidity being missed?
Maser JD, Kaelber C, Weise RE. International use and attitudes toward
growing consensus in psychiatric classification.
J Abnorm Psychol.
Widiger TA, Frances AJ, Pincus HA, et al, eds.
Vol 4. Washington, DC: American Psychiatric Press; 1998.
Stuart S, Pfohl B, Battaglia M, et al. The cooccurrence of
J Personal Disord.
Adler DA, Drake RE, Teague GB. Clinicians' practices in personality assessment: does gender influence the use of
O'Connor BP, Dyce JA. A test of models of personality disorder configuration.
J Abnorm Psychol.
Widiger TA, Costa PT Jr, McCrae RR. A proposal for Axis II: diagnosing personality disorders using the five-factor model. In: Costa PT Jr, Widiger TA, eds.
Personality Disorders and the Five-Factor Model of Personality.
2nd ed. Washington, DC: American Psychological Association; 2002:431-456.
Gunderson, JG. Diagnostic controversies. In: Tasman A, Riba MB, eds.
Review of Psychiatry.
Vol 11. Washington, DC: American Psychiatric Press; 1992:9-24.
Zimmerman M, Chelminski I. Generalized anxiety disorder in patients with major depression: is
s hierarchy correct?
Am J Psychiatry.
Tyrer P. The case for cothymia: an open verdict?
Br J Psychiatry.
Gude T, Vaglum P. One-year follow-up of patients with cluster C personality disorders: a prospective study comparing patients with "pure" and comorbid conditions with cluster C, and "pure" C with "pure" cluster A or B conditions.
J Personal Disord.
Robins E, Guze SB. Establishment of diagnostic validity in psychiatric illness: its application to schizophrenia.
Am J Psychiatry.
First MB, Pincus HA, Levine JB, et al. Using clinical utility as a criterion for revising psychiatric diagnoses.
Am J Psychiatry.