Journal of Affective Disorders 55 (1999) 55–61www.elsevier.com/ locate / jad

Brief report

Outcome for adjustment disorder with depressed mood: comparisonwith other mood disorders

a , b c c d*Rick Jones , William R. Yates , Sally Williams , Miranda Zhou , Lisa HardmanaLaureate Psychiatric Research Center, 6655 South Yale Avenue, Tulsa, OK 74136, USA

bDepartment of Psychiatry, University of Oklahoma, College of Medicine, Tulsa Campus, Tulsa, OK, USAcLaureate Research Center, Tulsa, OK, USA

dTulsa University, Tulsa, OK, USA

Received 18 May 1998; accepted 28 August 1998

Abstract

Background: A review of the research literature on the diagnostic category of adjustment disorder indicates that itsconstruct validity has not been established. Nevertheless, the diagnosis is made frequently, with an estimated incidence of5–21% in psychiatric consultation services for adults. Methods: Retrospective data was used to evaluate the constructvalidity of the adjustment disorder diagnostic category. The data primarily consisted of SF-36 Health Status Surveyresponses by a large group of adult psychiatric outpatients before treatment and again six months after beginning treatment.Subjects were divided into five diagnostic groups, and MANOVA, MANCOVA and chi square were used to clarifyrelationships among diagnoses, sociodemographic data and SF-36 scores. Results: Diagnostic categories were significantlydifferent at baseline, but did not differ in terms of outcome at six-months follow-up. There was a significant genderdifference at baseline and a significant difference in gender distribution across diagnostic categories. Limitations: Structuredinterviews were not used for initial diagnoses, nor is there an estimate of the reliability of diagnoses among the clinicians.The patient attrition rate for six-months follow-up data was about 50%. Finally, patients received individualized treatment,with some patients receiving both medication and psychotherapy. Conclusions: Female patients were significantly morelikely to be diagnosed with major depression or dysthymia than with an adjustment disorder. Females were also more likelythan males to score lower on the mental health related scales of the SF-36 at admission. Patients diagnosed with anadjustment disorder scored higher on all SF-36 scales than did the other diagnostic groups at baseline and again at follow-up.There was no significant difference among diagnostic groups with regard to treatment outcome, suggesting that theadjustment disorder group can benefit as much as the other groups from treatment. 1999 Elsevier Science B.V. All rightsreserved.

Keywords: SF-36; Adjustment disorder; Treatment outcome

*Corresponding author. Tel.: 1 1-918-491-3704; fax: 1 1-918-491-5792.E-mail address: LPRC@saintfrancis.com (R. Jones)

0165-0327/99/$ – see front matter 1999 Elsevier Science B.V. All rights reserved.PI I : S0165-0327( 98 )00202-X

56 R. Jones et al. / Journal of Affective Disorders 55 (1999) 55 –61

Even in relatively recent literature (Despland et in a quality assurance program which consisted ofal., 1995; Kovacs et al., 1995) the diagnostic cate- completion of a baseline SF-36 (cf. below) and agory of adjustment disorder is treated as though it follow-up SF-36 at six months after admission. Formight be a mirage. There are relatively few studies administrative reasons, not all patients were ap-of the disorder (Andreasen and Wasek, 1980 An- proached, and of course not all patients agreed todreasen and Hoenk, 1982; Cantwell and Baker, 1989 participate. However, we have no evidence of aFabrega et al., 1987; Newcorn and Strain, 1992; selection bias in the kinds of patients who wereKovacs et al., 1994) and none that convey confi- asked to participate. In other words, aside from thosedence in establishing its construct validity. In the last patients who declined to participate, we believe that25 years, there have been fewer than 25 articles the sampling was random. Slightly less than 50% ofdealing with the subject. In contrast, the estimated patients admitted from late 1995 through the end ofincidence of the disorder is high, ranging from 5– 1997 agreed to participate, and again slightly less21% in psychiatric consultation services for adults than 50% of these patients completed the SF-36 at(Despland et al., 1995; Andreasen and Wasek, 1980; the six-month follow-up.Fabrega et al., 1987). Several of these studies A primary goal of the Laureate Outpatient Pro-(Fabrega et al., 1987, 1986) have found differences gram is reduction of patients’ emotional distress andbetween patients with adjustment disorder and those improvement in their occupational and social func-with a more specific disorder, another mood disorder, tioning. The quality assurance program was designedor those with no diagnosed disorder. Differences to measure the effectiveness of the clinic in accom-include severity of symptoms, psychosocial adapta- plishing this goal. The clinic is part of a largertion, and number and intensity of stressors. Further complex that includes inpatient and partial psychiat-evidence (Fabrega et al., 1986; Snyder et al., 1990) ric hospitalization. Laureate Psychiatric Clinic andsuggests that patients diagnosed with adjustment Hospital (LPCH) is part of a larger full servicedisorder are younger, more likely to be impulsive, to healthcare system, St. Francis Health System. Use ofabuse toxic substances, and to have a personality the clinical database for research purposes wasdisorder. With regard to treatment outcome, (Looney approved by the St. Francis Health System Institu-and Gunderson, 1978; Bronish, 1991; Greenberg et tional Review Board.al., 1995) patients with adjustment disorder require Licensed mental health professionals performless treatment, are able to return to work sooner, and diagnostic interviews. Five psychiatrists coordinateare less likely to manifest a recurrence of the the base line psychiatric evaluation, using the DSM-disorder. IV system of classification. The American Board of

Despite skepticism about the diagnostic category Psychiatry and Neurology certifies the psychiatrists.from a methodological perspective (e.g. lack of Up to three psychiatric disorders are classified atspecificity of symptoms, behavioral parameters, or admission and coded for clinical and billing pur-close links with environmental factors) there appear poses. The first diagnosis is considered to be clinical-to be practical, clinical reasons for maintaining the ly predominant.category. The present study is a continuation of this Subjects were classified into five diagnosticpragmatic prospectus on adjustment disorder, how- groups: (1) major depression, single and (2) re-ever, focusing only on depressive diagnoses and current episodes, (3) dysthymia, (4) depression NOS,paying special attention to treatment outcome data as and (5) adjustment disorder with depressed mood ora means of establishing clinical validity for diagnos- mixed anxiety and depressed mood. Sociodemo-tic categories. graphic data for these five groups may be found in

Table 1.

1. Methods, subject selection2. Measures

Beginning in the latter part of 1995, adults seekingtreatment at the Laureate Outpatient Psychiatric The primary psychometric measure for this studyClinic in Tulsa, Oklahoma were asked to participate was the 36-item Short-Form Health Status Survey

R. Jones et al. / Journal of Affective Disorders 55 (1999) 55 –61 57

Table 1

Major depression Dysthymia Depression AdjustmentNOS disorder

Single RecurrentN 5 339 N 5 158 N 5 68 N 5 148 N 5 167

Age 38.10 65 36.38 39.99 31(SD) (11.27) (84) (39) (38) (12.00)Female* 241 117 50 92 97(N, %) (71.1%) (74.1%) (73.5%) (62.2%) (58.1%)Marital Status(N, %) 92 37 21 38 37Single, never married (27.1%) (23.4%) (30.9%) (25.7%) (22.2%)Married 191 88 35 83 99

(56.3%) (55.7%) (51.5%) (56.1%) (59.3%)Divorced/separated 55 33 12 27 30Widowed (16.3%) (20.9%) (17.6%) (18.2%) (18%)Race 326 151 65 142 160(N, %) Caucasian (96.4%) (96.2%) (97%) (95.9%) (95.8%)Employed** 277 123 56 115 146(N, %) (81.7%) (77.8%) (82.4%) (77.7%) (87.4%)a 2X (4,N 5 1281) 5 29.04, P , 0.001.b ‘Employed’ category includes individuals identified as students or housewives.

(SF-36). This survey is a self-report measure that tionnaires and those who did not, a discriminatewas adapted from the Medical Outcomes Study long- analysis was performed on the following variables:form measure. This measure is considered a measure gender, age, marital status, employment status, andof health status and functioning, well being and admission scores on the PCS and MCS. Employmentquality of life. The SF-36 contains eight multi-item status was coded as a dichotomous variable, eithersubscales including: social functioning, emotional employed or unemployed. Marital status was re-well-being, role limitation due to emotional health coded into two dichotomous variables, the first asproblems, energy or fatigue, physical functioning, married vs. not married, and the second as divorcedrole limitation due to physical functioning, freedom or separated vs. other. The diagnosis variable wasfrom pain, and general health perceptions. Two not included because a preliminary cross-tabulationsummary scores, the Physical Component Summary analysis resulted in no statistically significant differ-(PCS) and the Mental Component Summary (MCS) ence between responders and non-responders acrossare derived from the eight subscales. The SF-36 has the five diagnostic categories.been used in a variety of studies of the general Descriptive statistics were used to compare thepopulation as well as medical and psychiatric popu- socio-demographic profiles of the five diagnosticlations. Empirical studies support the construct va- category groups (Table 1). Baseline SF-36 scoreslidity of all the summary scales and subscales. Scale were contrasted by diagnosis by using MANOVAvalues for the subscales range from 0 (poorest (Table 2). A MANOVA was performed on the time-health) to 100 (optimal health). Scale values for the two SF-36 scores (Table 3). Then a MANCOVA wassummary scales are calculated as T scores. performed on those same scores using the time-one

scores as covariates. Finally, because of noticeablegender differences on the mental health related scale

3. Statistical analysis scores, a second MANOVA and MANCOVA wereperformed, using gender as a blocking variable, to

In order to detect possible differences between explore the possibility of a diagnosis x genderpatients who completed six-month follow-up ques- interaction. Additionally, a chi-square was performed

58 R. Jones et al. / Journal of Affective Disorders 55 (1999) 55 –61

Table 2Baseline SF-36 scores for each diagnostic group

SF-36 General Major depression Dysthymia Depression Adjustmentscale US population NOS disorder

Single RecurrentN 5 2474 N 5 314 N 5 144 N 5 62 N 5 135 N 5 157

Physical M 5 84.15 73.60 73.90 76.52 74.73 88.58functioning SD 5 23.28 (28.40) (28.15) (23.44) (30.65) (18.45)

Bodily pain 75.15 56.88 57.58 61.53 60.45 77.0623.69 (29.61) (28.82) (26.76) (29.64) (24.09)

Role- 80.96 52.42 51.22 50.81 59.63 78.98physical 34.00 (43.06) (44.18) (41.23) (42.54) (34.50)

General 71.95 56.44 52.25 54.23 60.61 74.19health 20.34 (25.27) (23.71) (24.07) (23.50) (21.39)

Vitality 60.86 24.46 21.01 29.38 29.78 42.1220.96 (19.87) (18.43) (20.60) (22.76) (21.35)

Role- 81.26 19.69 18.06 26.34 29.38 42.04emotional 33.04 (30.77) (33.42) (37.27) (38.21) (41.27)

Social 83.28 34.75 32.21 40.12 42.59 53.98functioning 22.69 (24.82) (23.95) (27.55) (25.57) (26.06)

Mental 74.74 31.87 30.63 37.23 39.01 47.48health 18.05 (19.72) (17.38) (22.13) (20.18) (20.23)

Physical M 5 50.00 49.03 48.75 48.77 49.60 56.45component SD 5 10.00 (13.04) (12.97) (11.13) (13.39) (10.08)

asummary

Mental 50.00 21.19 20.00 24.86 25.58 32.18component 10.00 (11.94) (11.02) (13.61) (12.88) (13.29)summary

MANOVA

Wilks’ F Hypothesis df Error df Significancelambda

Diagnostic category 0.846 44.26 32 2033.41 0.001Gender 0.997 2.29 8 795 0.02Gender 3 diagnosis 0.972 0.70 32 2933.41 0.894a Component summary scores use T scores.

to determine whether there was a relationship be- baseline, patients’ average score on the MCS wastween gender and diagnostic category (without re- 24.99 and their average age was 37.90; at follow-upgard to SF-36 scores). the average score on the MCS was 23.16 and the

average age was 38.93. The MANOVA on the eightSF-36 scales at baseline indicated that the five

4. Results diagnostic categories were significantly different(P , 0.001). Deviation contrasts indicated that both

The discriminate analysis resulted in statistically major depressive categories and the adjustmentsignificant differences between patients who com- disorder categories differed significantly from thepleted the six-month follow-up and those who did overall mean for all categories. The data in Table 2not on two of the variables: age and MCS. At indicate that patients diagnosed with major depres-

R. Jones et al. / Journal of Affective Disorders 55 (1999) 55 –61 59

Table 3Six-month follow-up SF-36 scores for each of the diagnostic groups before the variance attributable to time-one scores has been removed

SF-36 General Major depression Dysthymia Depression Adjustmentscale US population NOS disorder

Single RecurrentN 5 2474 N 5 152 N 5 79 N 5 32 N 5 57 N 5 77

Physical M 5 84.15 73.04 71.52 75.16 74.63 90.39functioning SD 5 23.28 (30.13) (28.08) (23.81) (29.20) (19.65)

Bodily 75.15 59.30 59.86 62.62 55.88 81.62pain 23.69 (30.50) (28.78) (23.81) (29.96) (23.40)

Role- 80.96 54.11 58.86 60.16 56.14 85.71physical 34.00 (42.72) (44.22) (44.39) (42.60) (29.08)

General 71.95 59.09 55.85 59.63 59.60 79.09health 20.34 (24.25) (25.06) (28.20) (23.82) (19.00)

Vitality 60.86 40.91 36.25 42.97 40.26 59.4220.96 (25.77) (25.37) (28.45) (24.54) (21.54)

Role- 81.26 47.81 42.41 53.12 48.54 69.70emotional 33.04 (41.62) (40.55) (43.88) (43.68) (37.54)

Social 83.28 57.48 52.37 63.67 57.02 78.08functioning 22.69 (31.13) (28.87) (28.99) (29.70) (24.52)

Mental 74.74 54.99 52.90 55.38 54.47 70.39health 18.05 (25.21) 21.37) (27.86) (24.09) (18.82)

Physical M 5 50.00 45.46 45.85 46.63 45.56 54.45component SD 5 10.00 (12.84) (13.76) (11.41) (11.95) (9.57)summary*

Mental 50.00 35.45 32.70 37.07 35.49 46.55component 10.00 (16.53) (14.73) (17.39) (16.08) (12.87)summary

MANCOVA

Wilks’ lambda F Hypothesis df Error df Significance

Diagnostic category 0.943 0.644 32 1277.58 0.939Gender 0.975 1.11 8 346 0.356Gender 3 diagnosis 0.935 0.731 32 1277.58 0.864

sion, single episode or recurrent, score lower on all the time-two (6-month follow-up) SF-36 scoreseight scales than do patients in the other depressive indicated that the five diagnostic categories werediagnostic categories. On the other hand, patients again significantly different (P , 0.001). Deviationdiagnosed with an adjustment disorder that includes contrasts indicated that the recurrent major depres-a depression component tend to score significantly sion and the adjustment disorder categories eachhigher on the eight scales than do the other depres- differed significantly from the overall mean for allsive patients. In fact, the physical health scores for categories. The time-two scores are presented inthe adjustment disorder patients are essentially the Table 3. A MANCOVA was performed on the time-same as those of the general U.S. population. No two (6-month follow-up) SF-36 scale scores for twostatistically significant differences were found be- reasons: to avoid the problem of the statisticaltween the dysthymia and depression, NOS groups on unreliability of change scores, and secondly toany of the eight SF-36 scale scores, even after post compare the five diagnostic groups from the samehoc comparisons were examined. The MANOVA on baseline score. The MANCOVA on the eight SF-36

60 R. Jones et al. / Journal of Affective Disorders 55 (1999) 55 –61

scales resulted in no statistically significant differ- time period, 63% were female. When this group wasences among any of the five diagnostic groups. restricted to those outpatients with depression-related

Finally, a MANOVA on the baseline SF-36 scores diagnoses (the five categories in this investigation),that included gender as a blocking variable showed 68% were female. Furthermore, within the fivethat there was a statistically significant difference diagnostic categories, 73% of those patients sufferingbetween males and females on baseline scores (P , from a major depressive disorder were female, while0.05), but that there was no significant interaction 57% of those suffering from an adjustment disorderbetween gender and diagnostic category. However, a were female. The noticeably greater proportion ofchi-square done on gender and diagnostic category females as psychiatric outpatients, as patients suffer-(without regard to SF-36 scores) indicated a statisti- ing from depression in particular, and finally ascally significant relationship between these two patients suffering from the more severe kinds ofvariables. A MANCOVA on the follow-up SF-36 depression is consistent with a line of publishedscale scores, with time-one scores as covariates and research (e.g. Akiskal, 1989); Perugi et al. (1990)gender as a blocking variable, resulted in no signifi- have demonstrated a greater degree of depressivecant difference between males and females with temperament in women and Nolen-Hoeksema (1987)regard to treatment outcome. has argued that women’s response style to depression

may exacerbate and prolong the episode. She hasdone a series of studies (Morrow and Nolen-Hoek-

5. Conclusions sema, 1990; Nolen-Hoeksema, 1991; Nolen-Hoek-sema et al., 1992, 1993; Lyubomirsky and Nolen-

The SF-36 data at baseline and at the six-month Hoeksema, 1993; Nolen-Hoeksema and Girgus,follow-up both support the separate construct validity 1994; Lyubomirsky and Nolen-Hoeksema, 1995)of the adjustment disorder diagnostic category. Both arguing for the importance of response style toMANOVAS were significant (P , 0.001), and in depression and the prevalence of a ruminative re-both analyses, deviation contrasts indicated that the sponse style in women after the age of 15. In ouradjustment disorder group scores were significantly investigation, female patients scored significantlydifferent (P , 0.05) from the overall mean for all lower than males on the baseline SF-36. There wasfive categories. Furthermore, the results of the MAN- also a significant gender by diagnosis difference, asCOVA indicating that there were no differences in mentioned above, but there was no significant gendersix-month outcome scores among the five diagnostic by diagnosis interaction with regard to baseline orcategories, when the variance attributable to baseline follow-up SF-36 scores. Nor did females differscores was eliminated, actually adds further support significantly from males on the follow-up SF-36,to the separateness of the adjustment disorder cate- suggesting that the female patients who participatedgory. These patients, even though they scored sig- in the study and who cooperated with the follow-upnificantly higher at the time of admission than did showed greater improvement in self-reported symp-patients in the other categories, still manifested tom status and functioning than did males.substantial improvement in response to treatment Although the results of this data analysis offerafter six months. Their improvement is equivalent to support for the existence of a separate diagnosticthat of patients with MDD (Major Depressive Disor- category for adjustment disorder, there are severalder) or dysthymia. Because these patients can dem- limitations that must be considered. Althoughonstrate substantial improvement in symptom status licensed mental health professionals, primarily boar-and functioning, they are surely deserving of atten- d-certified psychiatrists did initial diagnostic evalua-tion and treatment. tions of these patients, structured interviews were not

An interesting finding in this dataset is the gender used, nor is there an estimate of the reliability ofdifference. Of individuals seeking treatment at the diagnoses among these clinicians. Secondly, ourLPCH Outpatient Clinic for the last 2 years, 63% patient attrition rate was about 50%. Althoughwere female. Of outpatients agreeing to participate in discriminate analysis has found few differencesthe outcome assessment program during that same between those who gave us follow-up SF-36 data

R. Jones et al. / Journal of Affective Disorders 55 (1999) 55 –61 61

Snyder, S., Strain, J.J., Wolf, D., 1990. Differentiating majorand those who didn’t, we cannot rule out thedepression from adjustment disorder with depressed mood inpossibility of a systematic bias in our outcome data.the medical setting. Gen. Hosp. Psychiatry 12, 159–165.

Finally, each patient receives an individualized treat- Looney, J.G., Gunderson, E., 1978. Transient situational distur-ment, with some patients receiving both medication bances: Course and outcome. Am. J. Psychiatry 135, 660–663.and psychotherapy. Additional studies with reliabili- Bronish, T., 1991. Adjustment reactions: A long-term prospective

and retrospective follow-up of former patients in a crisisty checks on diagnosis, with a lower attrition rateintervention ward. Acta Psychiatr. Scand. 84, 86–93.and with a standardized treatment protocol are

Greenberg, W.M., Rosenfeld, D.N., Ortega, E.A., 1995. Adjust-needed to confirm our findings. ment disorder as an admission diagnosis. Am. J. Psychiatry

152, 459–461.Akiskal, H.S., 1989. New insights into the nature and hetero-

geneity of mood disorders. J. Clin. Psychiatry 50 (suppl.),References6–10.

Perugi, G., Musetti, L., Simoni, E., Piagentini, F., Cassano, G.B.,Despland, J.N., Monod, L., Ferrero, F., 1995. Clinical relevance of Akiskal, H.S., 1990. Gender-mediated clinical features of

adjustment disorder in DSM-III-R and DSM-IV. Comprehen- depressive illness: the importance of tempermental differences.sive Psychiatry 36, 454–460. Br. J. Psychiatry 157, 835–841.

Kovacs, M., Ho, V., Pollock, M., 1995. Criterion and predictive Nolen-Hoeksema, S., 1987. Sex differences in unipolar depres-validity of the diagnosis of adjustment disorder: A prospective sion: evidence and theory. Psychol. Bull. 101, 259–282.study of youths with new-onset insulin-dependent diabetes Morrow, J., Nolen-Hoeksema, S., 1990. Effects of responses tomellitus. Am. J. Psychiatry 152, 523–528. depression on the remediation of depressive affect. J. Pers. Soc.

Andreasen, N.C., Wasek, P., 1980. Adjustment disorders in Psychol. 58 (3), 519–527.adolescents and adults. Arch. Gen. Psychiatry 37, 1166–1170. Nolen-Hoeksema, S., 1991. Responses to depression and their

Andreasen, N.C., Hoenk, P.R., 1982. The predictive value of effects on the duration of depressive episodes. J. Abnorm.adjustment disorders: A follow-up study. Am. J. Psychiatry Psychol. 100 (4), 569–582.139, 584–590. Nolen-Hoeksema, S., Girgus, J.S., Seligman, M.E., 1992. Predic-

Cantwell, D.P., Baker, L., 1989. Stability and natural history of tors and consequences of childhood depressive symptoms: aDSM-III childhood diagnoses. J. Am. Acad. Child Adolesc. 5-year longitudinal study. J. Abnorm. Psychol. 101 (3), 405–Psychiatry 28, 691–700. 422.

Fabrega, H., Mezzich, J.E., Mezzich, A.C., 1987. Adjustment Nolen-Hoeksema, S., Morrow, J., Fredrickson, B.L., 1993. Re-disorder as a marginal or transitional illness category in DSM- sponse styles and the duration of episodes of depressed mood.III. Arch. Gen. Psychiatry 44, 567–572. J. Abnorm. Psychol. 102 (1), 20–28.

Newcorn, J.H., Strain, J., 1992. Adjustment disorder in children Lyubomirsky, S., Nolen-Hoeksema, S., 1993. Self-perpetuatingand adolescents. J. Am. Acad. Child Adolesc. Psychiatry 31, properties of dysphoric rumination. J. Pers. Soc. Psychol. 65318–326. (2), 339–349.

Kovacs, M., Gatsonis, C., Pollock, M., Parrone, P., 1994. A Nolen-Hoeksema, S., Girgus, J.S., 1994. The emergence of gendercontrolled prospective study of DSM-III adjustment disorder in differences in depression during adolescence. Psychol. Bull.childhood: Short-term prognosis and long-term predictive 115 (3), 424–443.validity. Arch. Gen. Psychiatry 51, 535–541. Lyubomirsky, S., Nolen-Hoeksema, S., 1995. Effects of self-

Fabrega, H., Mezzich, J.E., Mezzich, A.C., Coffman, G.A., 1986. focused rumination on negative thinking and interpersonalDescriptive validity of DSM-III depressions. J. Nerv. Ment. problem solving. J. Pers. Soc. Psychol. 69 (1), 176–190.Dis. 174, 573–584.