DEPRESSION AND ANXIETY 00:1–5 (2012)

Research ArticleWHY DO SOME DEPRESSED OUTPATIENTS WHO ARENOT IN REMISSION ACCORDING TO THE HAMILTON

DEPRESSION RATING SCALE NONETHELESS CONSIDERTHEMSELVES TO BE IN REMISSION?

Mark Zimmerman, M.D.,1,2∗ Jennifer Martinez, B.A.,1,2 Naureen Attiullah, M.D.,1,2 Michael Friedman,M.D.,1,2 Cristina Toba, M.D.,1,2 and Daniela A. Boerescu, M.D.1,2

Background: In treatment studies of depression, remission is typically definednarrowly—based on scores on symptom severity scales. Patients treated in clini-cal practice, however, define the concept of remission more broadly and considerfunctional status, coping ability, and life satisfaction as important indicators ofremission status. In the present report from the Rhode Island Methods to Im-prove Diagnostic Assessment and Services (MIDAS) project, we examined howmany mildly symptomatic depressed patients in ongoing treatment who did notscore in the remission range on the 17-item Hamilton Depression Rating Scale(HAMD) nonetheless considered themselves to be in remission from their de-pression. Among the mildly symptomatic HAMD nonremitters, we compared thedemographic and clinical characteristics of patients who did and did not con-sider themselves to be in remission. Methods: We interviewed 274 psychiatricoutpatients diagnosed with DSM-IV major depressive disorder who were in on-going treatment. The patients completed measures of psychosocial functioningand quality of life. Results: Approximately one-quarter of the patients scoring8–12 on the HAMD considered themselves to be in remission. Compared to pa-tients who did not consider themselves to be in remission, the remitters reportedsignificantly better quality of life, less functional impairment due to depression,higher positive mental health scores, and better coping ability. Discussion: Somepatients who do not meet symptom-based definitions of remission nonethelessconsider themselves to be in remission. The findings raise caution in relying ex-clusively on symptom-based definitions of remission to guide treatment decisionmaking in clinical practice. Depression and Anxiety 00:1–5, 2012. C© 2012Wiley Periodicals, Inc.

Key words: depression; remission; Hamilton Rating Scale for Depression

1Department of Psychiatry and Human Behavior, Brown Medi-cal School, Providence, Rhode Island2Department of Psychiatry, Rhode Island Hospital, Providence,Rhode Island

Conflicts of interest. None.

Funding/support: The research was supported, in part, by Eli LillyUSA, LLC.

∗Correspondence to: Mark Zimmerman, Bayside Medical Center,235 Plain Street, Providence, RI 02905.E-mail: mzimmerman@lifespan.org

Experts recommend remission as the primary goal inthe treatment of depression.[1–9] The implication of rec-ommendations to “treat till remission” is that treatmentshould be modified until remission is achieved.

Received for publication 15 December 2012; Revised 2 July 2012;Accepted 7 July 2012

DOI 10.1002/da.21987Published online in Wiley Online Library (wileyonlinelibrary.com).

C© 2012 Wiley Periodicals, Inc.

2 Zimmerman et al.

A potential problem with this recommendation is thattreatment modifications are based on a definition ofremission determined by a score on a symptom scale.Although there have been differences in operationaldefinitions of remission in acute treatment studies ofdepression,[10, 11] at their core, these definitions havebeen symptom-based and therefore narrow in scope.There is little data to suggest that these symptom-baseddefinitions of remission used in controlled outcome stud-ies adequately reflect the perspectives of depressed pa-tients receiving treatment in routine clinical settings. Infact, the amelioration or elimination of depression symp-toms, while an important goal, is not necessarily the pri-mary outcome that depressed patients wish to achievefrom treatment.[12] The three factors most frequentlyjudged to be very important in determining remissionfrom depression were the presence of features of posi-tive mental health such as optimism and self-confidence,a return to one’s usual, normal self, and a return to usuallevel of functioning. The current definitions of remis-sion used in clinical trials may therefore not adequatelyreflect the perspectives of patients treated in routine clin-ical practice.

Two types of discordance can occur between re-searchers’ and patients’ perceptions of remission—research-defined remission that is not confirmed by pa-tients’ attestation, and patients’ perception of being inremission despite not scoring in the remission on a stan-dardized symptom scale. In a previous report, we ex-amined how many patients who scored in the remis-sion range on the Hamilton Depression Rating Scaledid not consider themselves to be in remission, andamong the HAMD remitters, we compared the demo-graphic and clinical characteristics of patients who didand did not consider themselves to be in remission.[13]

We found that approximately half of the patients scoringin the remission range on the HAMD did not considerthemselves to be in remission, and the self-describedremitters had significantly lower levels of depressiveand anxious symptoms, better quality of life, less func-tional impairment due to depression, had higher pos-itive mental health scores, and reported better copingability.

In the present report, we turn to the other type ofdiscordance between research and patient evaluations ofremission—patients who fail to meet the HAMD defi-nition of remission but who consider themselves to bein remission. Our clinical experience indicates that somepatients who have improved but continue to experiencemild symptom levels, and therefore do not meet suchsymptom-based remission definitions, nonetheless con-sider themselves to be doing well and do not wish tochange their treatment. Thus, while they might not scorein the remission range on the HAMD, they nonethe-less consider themselves to be doing well enough sothat they are not interested in modifying treatment.These research findings and clinical observations leadto the present study from the Rhode Island Methodsto Improve Diagnostic Assessment and Services (MI-

DAS) project in which we examined how many mildlysymptomatic depressed patients considered themselvesto be in remission and then compared the patients whodid and did not consider themselves to be in remis-sion. We examined the same set of variables as ourprevious report and hypothesized that among patientswho would not be considered to be in remission, ac-cording to the HAMD those who nonetheless consid-ered themselves to be in remission would have lowerlevels of symptoms, more features of positive mentalhealth, better psychosocial functioning, greater life sat-isfaction, and better ability to cope with daily stressthan patients who did not consider themselves to be inremission.

METHODThe sample consisted of 274 psychiatric outpatients who were being

treated for DSM-IV major depressive disorder in the Rhode IslandHospital Department of Psychiatry outpatient practice. The RhodeIsland Hospital outpatient group predominantly treats individuals withmedical insurance on a fee-for-service basis, and it is distinct fromthe hospital’s outpatient residency training clinic that predominantlyserves lower income, uninsured, and medical assistance patients. Forapproximately half of the patients, the diagnosis of major depressivedisorder was based on the Structured Clinical Interview for DSM-IV(SCID),[14] whereas the other patients were diagnosed on the basis ofan unstructured clinical interview. The sample included 87 (31.8%)men and 187 (68.2%) women who ranged in age from 19 to 80 years(M = 49.0, SD = 13.9). The Rhode Island Hospital institutional reviewcommittee approved the research protocol, and all patients providedinformed, written consent.

The patients were rated by the authors on the 17-item HAMD andDSM-IV Global Assessment of Functioning (GAF) Scale. In a previousstudy, we found high inter-rater reliability when rating the HAMD(ICC = .97).[15] Patients completed several self-report scales includinga demographic form, the psychosocial functioning and quality-of-lifesubscales of the Diagnostic Inventory of Depression (DID),[16] andthe Remission from Depression Questionnaire (RDQ).[17]

The demographic form included a question regarding the patient’sperception of whether they were currently in remission from depres-sion (0 = yes, 1 = no). The term remission was not defined for patients;thus, they answered the question based on their personal conceptual-ization of the concept.

The Diagnostic Inventory for Depression (DID)[16] is a self-reportscale designed to assess the DSM-IV symptom inclusion criteria for amajor depressive episode, assess psychosocial impairment due to de-pression, and evaluate subjective quality of life. The six-item psychoso-cial functioning subscale assesses the amount of difficulty symptomsof depression have caused in usual daily responsibilities, relationshipswith significant others such as spouse, relationships with close familymembers, relationships with friends, participation in leisure activities,and overall level of functioning. Items are rated on a 5-point Likertscale (0 = no difficulty; 4 = extreme difficulty). The quality-of-life sub-scale assesses satisfaction with the same areas covered by the psychoso-cial functioning scale as well as global satisfaction with mental healthand physical health. Items are rated on a 5-point Likert scale (0 = verysatisfied; 4 = very dissatisfied). The DID quality-of-life and psychoso-cial impairment subscales achieved high levels of internal consistencyand test–retest reliability.

The RDQ consists of 41 items assessing multiple components ofremission such as positive mental health, symptom levels, and cop-ing ability.[17] The items refer to the prior week, and are rated on a

Depression and Anxiety

Research Article: Why Do Some Depressed Outpatients Consider Themselves to be in Remission? 3

3-point rating scale (not at all or rarely true; sometimes true; oftenor almost always true). In the present study, we examined the positivemental health and coping subscales, both of which have high internalconsistency and test–retest reliability.

DATA ANALYSISRemission on the HAMD is most commonly defined as a score of

7 or less.[18] We therefore focused on patients scoring 8–12 on the17-item HAMD, and we compared patients who did and did not con-sider themselves to be in remission. The independent variables wereexamined categorically as well as continuously. We a priori defined asindicators of clinically significant functional impairment scores of 2 orhigher indicating at least moderate impairment. A score of 3 or higheron the quality-of-life items indicated dissatisfaction in that domain.t-tests were used to compare groups on continuously distributed vari-ables, and chi-square tests were used to compare categorical variables.

RESULTSThe mean score on the 17-item HAMD for the entire

sample was 8.6 (SD = 6.9). Approximately a quarter ofthe sample scored 8–12 on the HAMD (n = 64, 23.4%).Self-reported remission status was missing for one pa-tient. Slightly more than a quarter of the patients scor-ing 8–12 on the HAMD considered themselves to be inremission (n = 18, 28.6%).

In the sample scoring 8–12 on the HAMD, we com-pared the 18 patients who did and the 45 patients whodid not consider themselves to be in remission. Therewas no difference between groups in gender (61.1% vs.66.7%, χ2 = 0.18, n.s.) or age (46.7 ± 12.3 vs. 48.7 ±17.0, t = 0.4, n.s.) The HAMD scores were similar inthe patients who did and did not consider themselves tobe in remission (9.7 ± 1.6 vs. 9.8 ± 1.3, t = 0.1, n.s.).

The data in Table 1 shows that across all functionaldomains the self-described remitters reported less func-tional impairment due to depression. The remitters weresignificantly less likely to report moderate functional im-pairment in at least one specific area (44.4% vs. 80.0%,χ2 = 7.7, P < .01), as well as significantly less likelyto report moderate impairment on the global rating of

TABLE 1. Psychosocial functioning (PF) in depressedoutpatients with mild symptoms according to the17-item Hamilton Rating Scale for Depression who doand do not consider themselves to be in remission

Self-reported Self-reported notPF domain, mean remission in remission(SD) (N = 18) (N = 45) t-Value

Work performance 0.4 (0.7) 1.9 (1.2) 6.5***

Marital relationship 0.6 (0.9) 1.8 (1.4) 3.5***

Family relationships 0.6 (0.9) 1.3 (1.2) 2.5**

Friendships 0.7 (1.0) 1.3 (1.1) 2.1*

Leisure 1.1 (1.2) 2.0 (1.4) 2.3*

Global rating of 1.2 (1.0) 2.4 (1.0) 4.1***

impairment*

P ≤ .05; **P ≤ .01; ***P ≤ .001.

TABLE 2. Quality of life (QOL) in depressedoutpatients with mild symptoms according to the17-item Hamilton Rating Scale for Depression who doand do not consider themselves to be in remission

Self-reported Self-reported notQOL domain, mean remission in remission(SD) (N = 18) (N = 45) t-Value

Work performance 1.1 (1.0) 2.1 (1.0) 3.8**

Marital relationship 0.9 (0.8) 1.8 (1.3) 2.9*

Family relationships 1.1 (1.1) 1.5 (1.2) 1.2Friendships 1.4 (1.1) 1.6 (1.1) 0.5Leisure 1.4 (1.0) 2.2 (1.1) 2.8*

Mental health 1.3 (1.1) 2.7 (1.1) 4.3**

Physical health 1.5 (1.2) 2.3 (1.1) 2.7*

Global rating of life 1.4 (0.9) 2.4 (0.9) 3.9**

satisfactionGlobal rating of quality 1.3 (0.8) 2.1 (0.7) 3.9**

of life*

P ≤ .01; **P ≤ .001.

impairment (22.2% vs. 75.6%, χ2 = 15.3, P < .001). Re-sults were similar for the quality-of-life analyses. That is,compared to patients who did not consider themselvesto be in remission, the remitters reported significantlybetter quality of life across all domains (Table 2). More-over, the remitters were significantly less likely to be dis-satisfied in at least one specific area (50.0% vs. 80.0%,χ2 = 5.7, P < .05), although not significantly less likelyto report dissatisfaction on the global rating of qualityof life (11.1% vs. 31.1%, χ2 = 2.7, n.s.). Remitters weresignificantly less likely to report dissatisfaction in theirmental health (16.7% vs. 62.2%, χ2 = 10.7, P = .001).Finally, the remitted patients had higher scores on thepositive mental health subscale of the RDQ (15.5 ± 5.5vs. 8.0 ± 4.6, t = –5.4, P < .001) and on the coping sub-scale of the RDQ (3.9 ± 1.5 vs. 2.2 ± 1.2, t = –4.8, P <.001).

DISCUSSIONIn clinical practice, recommendations to modify treat-

ment for depression if remission has not been attained,and the corollary to continue the present course of treat-ment if the remission threshold has been reached, shouldnot rely on a definition of remission that is based exclu-sively on symptom status. We consider remission to be abroader construct than symptom level, a construct thatincludes other indicators of clinical status such as func-tioning, quality of life, resiliency in coping with stress,and a general sense of well-being. In the present study,approximately one-quarter of the patients who did notscore in the remission range on the HAMD nonethe-less considered themselves to be in remission fromdepression. It is not surprising that compared to the pa-tients who did not consider themselves to be in remis-sion, the self-described remitters reported better func-tioning, better quality of life, better coping ability, and

Depression and Anxiety

4 Zimmerman et al.

better overall mental health. This is consistent with ourclinical experience and explains why some patients withlow levels of symptoms, in the so-called nonremissionrange of symptom severity, nonetheless are not inter-ested in modifications to their treatment. The resultsare also consistent with our report focusing on patientsscoring in the remission range of HAMD.[13]

The results of this study have implications regard-ing efforts to adopt measurement-based care in clini-cal practice.[19] Although we embrace the concept ofmeasurement-based care to guide treatment,[20, 21] webelieve that a focus on symptoms is overly narrow. Apatient may be minimally symptomatic but still function-ing well, coping well with stress, feeling optimistic aboutthe future, and feeling self-confident. Such patients willresist treatment modification. The clinical importanceof treating to remission is to reduce the likelihood of re-lapse; however, whether or not mildly symptomatic indi-viduals who report good functioning and coping abilityare as vulnerable to relapse as mildly symptomatic pa-tients with persistent impairment in functioning has notbeen studied. The bottom line is that the measurement-based care movement will need to carefully consider thedesired goals of treatment and how to operationalizethe concept of remission. We believe that the defini-tion of remission should include nonsymptom-relatedelements, and would expect a broader assessment includ-ing multiple domains related to remission would betterpredict relapse rates than measures based on symptomsalone.

Before concluding, the limitations of the study shouldbe recognized. First, we examined only the 17-item ver-sion of the HAMD. We focused on the 17-item HAMDbecause it is the most commonly used measure in an-tidepressant efficacy trials, and the cutoff used to defineremission has been generally accepted. We would an-ticipate that our findings would be similar in studies oflonger versions of the HAMD as well as other depressionseverity scales such as the Montgomery–Asberg Depres-sion Rating Scale. Second, all of the assessments werecross-sectional. A longitudinal assessment to ascertainpersistent improvement in the different domains wouldprovide valuable information. Third, self-perceived re-mission status was based on the patients response toa single question. We did not want to limit or influ-ence the patients’ conceptualization of the concept andtherefore did not define the term. This likely resulted inheterogeneity among patients in their definition of re-mission thereby potentially introducing error variance.However, despite such error variance, expected differ-ences were found between patients who did and did notconsider themselves to be in remission. Fourth, the sam-ple was drawn from a single large general adult outpa-tient private practice setting in which the majority of thepatients were white, female, and in their 30s and 40s.Generalizability to samples with different demographiccharacteristics needs to be demonstrated. Fifth, we didnot systematically record the treatments received by pa-tients. Patients received different medications, and a sub-

set of patients was receiving psychotherapy, thereby in-creasing the generalizability of the findings to routineclinical practice. However, we did not examine whethermodifications in treatment differentiated HAMD remit-ters who did and did not consider themselves to be inremission from depression. This would be useful to ex-amine in future studies. And sixth, we did not assessdiagnostic comorbidity. While we asked patients to indi-cate whether they considered themselves to be in remis-sion from depression, perhaps the patients who did notconsider themselves in remission were considering thesymptoms associated with comorbid psychiatric disor-ders. The issue of diagnostic comorbidity and its impacton defining remission may be less relevant to industry-sponsored treatment trials because patients with co-morbid conditions are often excluded.[22] However, inreal-world clinical practice, the impact of comorbidityon defining remission, and how this might impact onalgorithm-based treatment recommendations, warrantsstudy.

REFERENCES1. Ballenger J. Clinical guidelines for establishing remission

in patients with depression and anxiety. J Clin Psychiatry1999;60(Suppl 22):29–34.

2. Ferrier I. Treatment of major depression: is improvement enough?J Clin Psychiatry 1999;60(Suppl 6):10–14.

3. Nierenberg A, Wright E. Evolution of remission as the new stan-dard in treatment of depression. J Clin Psychiatry 1999;60:7–11.

4. Rush A, Crismon M, Toprac M, et al. Consensus guidelines inthe treatment of major depressive disorder. J Clin Psychiatry1998;59(Suppl 20):73–84.

5. Rush A, Trivedi M. Treating depression to remission. PsychiatrAnn 1995;25:704–709.

6. Stahl S. Why settle for silver, when you can go for gold? Responsevs. recovery as the goal of antidepressant therapy. J Clin Psychiatry1999;60:213–214.

7. Thase M. Evaluating antidepressant therapies: remission asthe optimal outcome. J Clin Psychiatry 2003;64(Suppl 13):18–25.

8. Moller HJ. Outcomes in major depressive disorder: the evolvingconcept of remission and its implications for treatment. WorldBiol Psychiatry 2008;9:102–114.

9. McIntyre R, Fallu A, Konarski J. Measurable outcomes in psy-chiatric disorders: remission as a marker of wellness. Clin Ther2006;28:1882–1891.

10. Prien RF, Carpenter LL, Kupfer DJ. The definition and opera-tional criteria for treatment outcome of major depressive disorder.Arch Gen Psychiatry 1991;48:796–800.

11. Rush AJ, Kraemer HC, Sackeim HA, et al. Report by the ACNPTask Force on response and remission in major depressive disor-der. Neuropsychopharmacology 2006;31:1841–1853.

12. Zimmerman M, McGlinchey J, Posternak M, et al. Howshould remission from depression be defined? The de-pressed patient’s perspective. Am J Psychiatry 2006;163:148–150.

13. Zimmerman M, Martinez J, Attiullah N, et al. Why do some de-pressed outpatients who are in remission according to the Hamil-ton Depression Rating Scale not consider themselves to be in re-mission? J Clin Psychiatry 2012;73:790–795.

Depression and Anxiety

Research Article: Why Do Some Depressed Outpatients Consider Themselves to be in Remission? 5

14. First MB, Spitzer RL, Gibbon M, Williams JBW. 1995. Struc-tured Clinical Interview for DSM-IV Axis I Disorders – Patientedition (SCID-I/P, version 2.0). New York: Biometrics ResearchDepartment, New York State Psychiatric Institute.

15. Zimmerman M, Posternak MA, Chelminski I. Heterogeneityamong depressed outpatients considered to be in remission.Compr Psychiatry 2007;48(2):113–117.

16. Zimmerman M, Sheeran T, Young D. The Diagnostic in-ventory for depression: a self-report scale to diagnose DSM-IV for major depressive disorder. J Clin Psychol 2004;60:87–110.

17. Zimmerman M, Galione J, Attiullah N, et al. Depressed pa-tients perspectives of two measures of outcome: the Quick In-ventory of Depressive Symptomatology (QIDS) and the Remis-sion from Depression Questionnaire (RDQ). Ann Clin Psychiatry2011;23:208–212.

18. Frank E, Prien R, Jarrett R, et al. Conceptualization and rationalefor consensus definitions of terms in major depressive disorder.Arch Gen Psychiatry 1991;48:851–855.

19. American Psychiatric Association. 2010. Practice Guideline forthe Treatment of Patients with Major Depressive Disorder.Washington, DC: American Psychiatric Association.

20. Zimmerman M, McGlinchey JB, Chelminski I. An inadequatecommunity standard of care: lack of measurement of outcomewhen treating depression in clinical practice. Prim Psychiatry2008;15:67–75.

21. Zimmerman M, Young D, Chelminski I, Dalrymple K. How canyou improve quality without measuring outcome? Getting fromhere to there. Prim Psychiatry 2010;17:46–53.

22. Zimmerman M, Mattia JI, Posternak MA. Are subjects in pharma-cological treatment trials of depression representative of patientsin routine clinical practice. Am J Psychiatry 2002;159:469–473.

Depression and Anxiety