Research Article

DEPRESSION AND ANXIETY 27 : 977–981 (2010)

VALIDITY OF A SIMPLER DEFINITION OF MAJORDEPRESSIVE DISORDER

Mark Zimmerman, M.D.,� Janine N. Galione, B.S., Iwona Chelminski, Ph.D., Diane Young, Ph.D.,Kristy Dalrymple, Ph.D., and Caren Francione Witt, Ph.D.

Background: In previous reports from the Rhode Island Methods to ImproveDiagnostic Assessment and Services project, we developed a briefer definition ofmajor depressive disorder (MDD), and found high levels of agreement betweenthe simplified and DSM-IV definitions of MDD. The goal of the present studywas to examine the validity of the simpler definition of MDD. We hypothesizedthat compared to patients with adjustment disorder, patients with MDD wouldbe more severely depressed, have poorer psychosocial functioning, have greatersuicidal ideation at the time of the intake evaluation, and have an increasedmorbid risk for depression in their first-degree family members. Methods: Wecompared 1,486 patients who met the symptom criteria for current MDDaccording to either DSM-IV or the simpler definition to 145 patients with acurrent diagnosis of adjustment disorder with depressed mood or depressed andanxious mood. Results: The patients with MDD were more severely depressed,more likely to have missed time from work due to psychiatric reasons, reportedhigher levels of suicidal ideation, and had a significantly higher morbid risk fordepression in their first-degree family members. Both definitions of MDD werevalid. Conclusions: The simpler definition of MDD was as valid as the DSM-IVdefinition. This new definition offers two advantages over the DSM-IVdefinition—it is briefer and therefore more likely to be recalled and applied inclinical practice, and it is free of somatic symptoms thereby making it easier toapply with medically ill patients. Depression and Anxiety 27:977–981, 2010.r 2010 Wiley-Liss, Inc.

Key words: major depressive disorder; diagnostic criteria; validity; depression

INTRODUCTIONThe DSM-IV criteria for major depressive disorder(MDD) require the presence of at least five symptomsfrom a list of nine, one of which must be either lowmood or loss of interest or pleasure in all, or almost all,usual activities. While the symptom inclusion criteriafor MDD have remained essentially the same for thepast 35 years, there are problems with their use. In asurvey of physicians’ reported use of the DSM-IVMDD criteria, Zimmerman and Galione[1] found thatapproximately one-quarter of experienced psychiatristsoften do not determine if the MDD criteria are metwhen diagnosing depression, and that the majority ofprimary care physicians usually do not determinewhether the MDD criteria are met when diagnosingdepression. The results of this survey were consistentwith studies which found that most nonpsychiatrist

physicians are unable to recall most of the MDDsymptom criteria.[2–6]

Published online 28 June 2010 in Wiley Online Library (wiley

onlinelibrary.com).

DOI 10.1002/da.20710

Received for publication 23 March 2010; Revised 8 April 2010;

Accepted 15 April 2010

�Correspondence to: Mark Zimmerman, Bayside Medical Building,

235 Plain Street, Providence, RI 02905.

E-mail: mzimmerman@lifespan.org

The authors report they have no financial relationships within the

past 3 years to disclose.

The Department of Psychiatry and Human Behavior, Rhode

Island Hospital, Brown Medical School, Providence, Rhode

Island

rr 2010 Wiley-Liss, Inc.

These findings raised concerns about the clinicalutility of the MDD criteria. We therefore developed asimpler definition of MDD consisting of only fivecriteria, at least three of which need to be present(including low mood or loss of interest) in order todiagnose MDD.[7] As shown in Table 1, a high level ofagreement between the simpler definition and theDSM-IV definition has been found in each of sixsamples.[7–9] The mean level of agreement between thetwo definitions across the studies was 94.6%.

No previous report examined the validity of thesimpler definition.[7,8] We argued that improvedclinical utility was sufficient to consider revising thediagnostic criteria. Nonetheless, improving the clinicalutility of the definition of MDD while sacrificingvalidity would argue against changing the diagnosticcriteria. Accordingly, in the present report from theRhode Island Methods to Improve Diagnostic Assess-ment and Services (MIDAS) project, we turned to theissue of the validity of the simpler definition of MDD.To test the validity of the simpler and DSM-IVdefinitions of MDD, we compared the patientsdiagnosed with MDD according to each algorithm topatients with adjustment disorder with depressed moodor depressed and anxious mood. Because patients withadjustment disorder with depressed mood also have adisturbance in mood, they are a particularly appro-priate comparison group for a study of the validity ofdefinitions of MDD. We hypothesized that comparedto patients with adjustment disorder, patients withMDD would be more severely depressed, have poorerpsychosocial functioning, have greater suicidal ideationat the time of the intake evaluation, and have anincreased morbid risk for depression in their first-degree family members. We predicted that both thesimpler and DSM-IV definitions of MDD would bevalid.

METHODS

The Rhode Island MIDAS project represents an integration ofresearch methodology into a community-based outpatient practiceaffiliated with an academic medical center.[10–12] To date, 2,900psychiatric outpatients have been evaluated with a semi-structured

diagnostic interview in the Rhode Island Hospital Department ofPsychiatry outpatient practice. In the present report, we focus on the1,486 patients who met the symptom criteria for current majordepression according to either DSM-IV or the simpler definition. Inaddition, as a comparison group to establish the validity of the twodefinitions of major depression, we included 152 patients with acurrent diagnosis of adjustment disorder with depressed mood ordepressed and anxious mood. Seven patients who met the criteria forboth adjustment disorder and the simpler definition of MDD wereincluded in the MDD group because this is consistent with the DSM-IVhierarchical relationship between MDD and adjustment disorder.This left 145 patients in the adjustment disorder group. Thus, thepresent sample consisted of a total of 1,631 psychiatric outpatientswith MDD or adjustment disorder. The data in Table 2 show that themajority of the patients were white (85.9%), female (64.9%), married(42.2%) or single (27.7%), and graduated high school (64.4%).

All patients were interviewed by a diagnostic rater who adminis-tered the Structured Clinical Interview for DSM-IV (SCID).[13]

Because we were interested in the psychometric performance of the

TABLE 1. Concordance between a simpler definition of major depressive disorder with the DSM-IV symptom criteriafor major depression in six samples

Samples NNo. of patients meeting

DSM-IV symptom criteria Sensitivity Specificity Overall agreement Kappa

Zimmerman et al.[7]

Derivation sample 805 436 92.7 94.8 93.7 0.87Cross-validation sample 789 466 92.3 96.6 94.0 0.88

Andrews et al.[9]

General population 10,641 339 92.9 99.8 94.6 0.93Zimmerman et al.[8]

Psychiatric outpatients 1,100 476 89.1 93.9 91.8 0.83Pathological Gamblers 210 59 89.8 96.0 94.3 0.86Candidates for bariatric surgery 1,200 31 80.6 99.6 99.1 0.82

TABLE 2. Demographic characteristics of 1,631psychiatric outpatients with major depressive disorder oradjustment disorder with depressed mood

Characteristic n %

GenderFemale 1,059 64.9Male 572 35.1

EducationLess than high school 167 10.2Graduated high school 1,050 64.4Graduated college or greater 414 25.4

Marital statusMarried 688 42.2Living with someone 93 5.7Widowed 32 2.0Separated 103 6.3Divorced 263 16.1Never married 452 27.7

RaceWhite 1,401 85.9Black 83 5.1Hispanic 56 3.4Asian 10 0.6Other 81 5.0

Age (years) M 5 39.3 SD 5 12.4

978 Zimmerman et al.

Depression and Anxiety

DSM-IV symptom criteria for MDD, we modified the SCID andeliminated the skip-out that curtails the depression module forpatients who did not report either depressed mood or loss of interestor pleasure. Thus, we inquired about all of the symptoms ofdepression for all patients. The Rhode Island Hospital institutionalreview committee approved the research protocol, and all patientsprovided informed, written consent.

As an ongoing part of the MIDAS project, joint-interviewdiagnostic reliability information was collected on 48 participants.The reliability coefficients of the symptoms of depression rangedfrom .54 to .94 (mean k 5 .80).

The interview also included items from the Schedule for AffectiveDisorders and Schizophrenia on the level of social functioning duringthe past 5 years, and the amount of time missing work due topsychiatric reasons during the past 5 years.

Family history diagnoses were based on information provided bythe patient. The interview followed the guide provided in the FamilyHistory Research Diagnostic Criteria[14] and assessed the presence orabsence of problems for a variety of psychiatric disorders, although inthe present report we focused on depression in the patients’ first-degree family members. Morbid risks were calculated using age-corrected denominators or bezugsiffers based on Weinberg’s shortermethod.[15] Thus, relatives over the age of risk for the particularillness were given a value of 1; those within the age for risk were givena value of 0.5, and those below it were given a value of 0. Limits forthe ages of risk for depression were 25 to 44 years based on thedistribution of ages of onset in our probands. Morbid risks werecompared using the w2 statistic.

DATA ANALYSIS

Previously, we developed a simpler definition of MDD exclusive ofsomatic symptoms with the goal of maximizing concordance with thecurrent DSM-IV definition.[7] (Technically, our previous and thecurrent research have focused on the symptom criteria of a majordepressive episode. For stylistic simplicity we refer to this as a simplerdefinition of MDD.) We approached the development of a newdefinition in six ways, each of which yielded comparable results(agreement rates with the original DSM-IV definition ranged from

92.6 to 95.4%). In a replication and extension study, we examined thesimplest of these definitions: at least three of the following fivesymptoms are present (low mood, loss of interest, guilt orworthlessness, impaired concentration or indecisiveness, and deathwishes or suicidal thoughts), one of which is low mood or loss ofinterest.[8] For both definitions of MDD, we compared the groups tothe patients with adjustment disorder. Categorical variables werecompared by the w2 statistic, or by Fisher’s Exact Test if the expectedvalue in any cell of a 2� 2 table was less than 5. Continuous variableswere compared by t-test.

RESULTSFor the entire group of 2,900 patients, the pre-

valence of current MDD was 48.5% based on theDSM-IV criteria and 46.9% based on the simplerdefinition. The overall level of agreement between thesimplified and DSM-IV definition of MDD was 92.9%(k5 0.86). The prevalence of adjustment disorder withdepressed mood or depressed and anxious mood was5.0% (n 5 145).

We compared the patients who met each definitionof MDD to the patients with adjustment disorder.There were no differences in the demographiccharacteristics between the two groups.

The data in Table 3 show that compared to thepatients with adjustment disorder the patients withMDD according to each definition were more likelyto have missed a month or more of work due topsychiatric reasons during the previous 5 years, moreseverely depressed at the time of the evaluation,reported higher levels of suicidal ideation, poorersocial support, and had a significantly higher morbidrisk for depression in their first-degree familymembers.

TABLE 3. Clinical characteristics of psychiatric outpatients with adjustment disorder and patients meeting criteria for asimpler definition of MDD and the DSM-IV MDD criteria

Adjustment disorder(n 5 145)

MDD-simpler definition(n 5 1,360)a

MDD—DSM-IV definition(n 5 1,406)b

Clinical global impression of depression severity,mean (SD)

1.8 (0.7) 3.1 (0.6)��� 3.2 (0.6)���

Severity of suicidal ideation, mean (SD)c 0.5 (1.0) 1.5 (1.3)��� 1.4 (1.3)���

Global assessment of functioning, mean (SD) 59.8 (8.2) 49.2 (8.3)��� 49.3 (8.3)���

Best social functioning past 5 years, N (%)c

Good or excellent 120 (82.8) 843 (62.0)��� 885 (63.0)���

Fair or worse 25 (17.2) 516 (38.0) 519 (37.0)Time out of work past 5 years, N (%)d

Virtually no time 88 (71.0) 348 (28.3)��� 355 (27.9)���

One month or more 36 (29.0) 882 (71.7) 916 (72.1)Family history of depression, N (morbid risk)e 61 (7.8) 916 (12.4)��� 952 (12.4)���

MDD, major depressive disorder.aMDD-Simpler definition versus adjustment disorder, ���Po.001.bMDD-DSM-IV definition versus adjustment disorder, ���Po.001.cRating of 0–6 from Schedule of Affective Disorders and Schizophrenia.dPatients who were not expected to work (e.g. because they were students, physically ill) were not included.eThe number of relatives at risk for the morbid risk calculations were 778.5 for adjustment disorder, 7,402.5 for simpler definition, and 7,657.5 forDSM-IV definition.

979Research Article: Validity of a Simpler Definition of Major

Depression and Anxiety

DISCUSSIONThe present findings are important for two reasons.

First, they demonstrate that a simpler definition ofMDD can be developed that is as valid as the DSM-IVdefinition. Second, the findings highlight the validity ofdistinguishing between MDD and adjustment disorder.

After eliminating the four somatic criteria from theDSM-IV definition of MDD leaving the five mood andcognitive features, a high level of concordance wasfound between this simpler definition of MDD withthe original DSM-IV classification. This new definitionoffers two advantages over the DSM-IV definition—it isbriefer and therefore more likely to be recalled andapplied in clinical practice, and it is free of somaticsymptoms thereby making it easier to apply withmedically ill patients.

As shown in Table 1, a high level of concordancebetween the simpler and DSM-IV definitions ofMDD has been found in six samples—three psychiatricoutpatient samples, a general population communitysample, a sample of gamblers who often were depres-sed but did not present with depression as theirprimary complaint, and a sample of obese subjectswho often had medical comorbidity. None of theprior studies, however, examined the validity of thesimpler definition of MDD. The present resultsindicate that MDD diagnosed according to eitherof the two definitions was distinguishable fromadjustment disorder.

In our previous reports on the correspondencebetween the simpler and DSM-IV definitions ofMDD, we discussed whether the current MDD criteriashould be changed in the absence of improved validity.The present findings do not provide evidence thatthe simpler definition is more valid than theDSM-IV definition. However, the clinical utility ofthe simpler definition is likely to be greater than theDSM-IV definition, and this might improve validity inthe clinical setting. Although a simpler definition ofMDD will not enhance validity in research studiesusing careful and thorough assessment proceduresbased on semi-structured diagnostic interviews, webelieve that in clinical practice a simpler set of criteriafor diagnosing MDD might improve validity becausewe suspect that MDD is sometimes underdiagnosed inmedically ill patients due to the uncertainty as towhether or not to count the somatic criteria, andsometimes overdiagnosed when clinicians, particularlynonpsychiatrist physicians, do not fully evaluate thediagnostic criteria and diagnose MDD when fewer thanthe minimum number of features are present. Theresults of the present study supported the validity of thedistinction between MDD and ‘‘subthreshold’’ variantsof depression such as adjustment disorder. If thisdistinction is not made in clinical practice because thecriteria are too long or complex to apply, then nomatter how valid the DSM-IV criteria may be inresearch settings their validity in clinical settings will be

compromised by failing to follow the DSM-IVdiagnostic algorithm. The validity of clinical diagnosisis likely to be improved if a briefer, more clinicallyuseful, albeit equally valid, set of criteria is adopted.

In deciding how to proceed in the next versions ofthe DSM and ICD, the conceptual and practicaladvantages of a briefer set of criteria that is easierto apply to all patients, particularly medically illpatients and patients seen in primary care, needs tobe weighted against the disadvantages of deviatingfrom tradition and the risk of overlooking symptomsthat are important to assess in depressed patientseven though they are no longer diagnostic criteria.Is a potential gain in clinical utility and clinical validity,in the context of data demonstrating equal, though notsuperior, validity when using research assessmentprocedures, sufficient to warrant criteria modification?

Both psychiatrists and primary care clinicians reportthat they often do not use the DSM-IV criteriawhen diagnosing depression.[1] There are likely multi-ple reasons why clinicians, particularly primary careclinicians, do not use the DSM-IV criteria for MDDwhen diagnosing depression, with the length andcomplexity of the criteria being only one reason. Itis possible that even with an abbreviated set ofdiagnostic criteria clinicians still will not formallyapply them but instead will continue to makenoncriteria-based gestalt judgments regarding thepresence or absence of depression. Perhaps, then, achange in diagnostic criteria based on clinical utilitygrounds should require a demonstration of improvedclinical utility. To date, no study has examined whethera simpler definition of MDD would, in fact, improvethe validity of diagnosis in clinical practice. Also, it ispossible that with the increased use of self-adminis-tered depression screening scales to assist withdiagnostic evaluations that the DSM-IV criteria willbe applied more faithfully, and a simpler definition willnot enhance validity.

A limitation of the present study is that it wasconducted in a single outpatient practice in which themajority of the patients were white, female, and hadhealth insurance. Replication of the results in sampleswith different demographic characteristics is war-ranted. Also, replication in a sample of medically illpatients is important because the elimination of thesomatic criteria might have its greatest influence inthese patients. Depressed patients seen in the specialtyhealth care sector may be more severely ill thanpatients seen in primary care, therefore replication ina primary care setting, where the level of severity maynot be as great, is warranted.

REFERENCES1. Zimmerman M, Galione J. Psychiatrists reported use of the

DSM-IV criteria for major depressive disorder. J Clin Psychiatry2010;71:235–238.

980 Zimmerman et al.

Depression and Anxiety

2. Krupinski J, Tiller J. The identification and treatment ofdepression by general practitioners. Aust N Z J Psychiatry2001;35:827–832.

3. Medow M, Borowsky S, Dysken S, et al. Internal medicalresidents’ ability to diagnose and characterize major depression.West J Med 1999;170:35–40.

4. Gerrity M, Cole S, Dietrich A, Barrett J. Improving therecognition and management of depression: is there a role forphysician education? J Fam Pract 1999;48:949–957.

5. Learman L, Gerrity M, Field D, et al. Effects of a depressioneducation program on residents’ knowledge, attitudes, andclinical skills. Obstet Gynecol 2003;101:167–174.

6. Rapp S, Davis K. Geriatric depression: physicians’ knowledge,perceptions, and diagnostic practices. Gerontologist 1989;29:252–257.

7. Zimmerman M, Chelminski I, McGlinchey JB, Young D.Diagnosing major depressive disorder X: can the utility of theDSM-IV symptom criteria be improved? J Nerv Ment Dis2006;194:893–897.

8. Zimmerman M, Galione JN, Chelminski I, et al. A simplerdefinition of major depressive disorder. Psychol Med2010;40:451–457.

9. Andrews G, Slade T, Sunderland M, Anderson T. Issues forDSM-V: simplifying DSM-IV to enhance utility: the case ofmajor depressive disorder. Am J Psychiatry 2007;164:1784–1785.

10. Zimmerman M. Integrating the assessment methods of researchersin routine clinical practice: the Rhode Island Methods toImprove Diagnostic Assessment and Services (MIDAS) project.In: First M, editor. Standardized Evaluation in Clinical Practice.Washington, DC: American Psychiatric Publishing, Inc.; 2003:29–74.

11. Posternak MA, Zimmerman M, Solomon DA. Integratingoutcomes research into clinical practice: a pilot study. PsychiatrServ 2002;53:335–336.

12. Zimmerman M, Mattia JI, Posternak MA. Are subjects inpharmacological treatment trials of depression representative ofpatients in routine clinical practice. Am J Psychiatry2002;159:469–473.

13. First MB, Spitzer RL, Gibbon M, Williams JBW. StructuredClinical Interview for DSM-IV Axis I Disorders—PatientEdition (SCID-I/P, version 2.0). New York: BiometricsResearch Department, New York State Psychiatric Institute;1995.

14. Endicott J, Andreasen N, Spitzer RL. Family History ResearchDiagnostic Criteria. 3rd ed. New York: Biometrics Research,New York State Psychiatric Institute; 1978.

15. Stromgren E. Statistical and genetic population studies withinpsychiatry: methods and principal results. Actualities Scientifi-ques et Industrelles. 1101 Congres International de Psychiatrie1950;6:155–157.

981Research Article: Validity of a Simpler Definition of Major

Depression and Anxiety