Journal of Affective Disorders 55 (1999) 45–49www.elsevier.com/ locate / jad

Brief report

Dysfunctional attitudes and poor problem solving skills predictqhopelessness in major depression

Barbara Cannon, Rosemarie Mulroy, Michael W. Otto, Jerrold F. Rosenbaum,*Maurizio Fava, Andrew A. Nierenberg

Massachusetts General Hospital and Harvard Medical School, WACC-815, 15 Parkman Street, Boston, MA 02114, USA

Received 6 May 1998; accepted 19 May 1998

Abstract

Background: Hopelessness is a significant predictor of suicidality, but not all depressed patients feel hopeless. If clinicianscan predict hopelessness, they may be able to identify those patients at risk of suicide and focus interventions on factorsassociated with hopelessness. In this study, we examined potential predictors of hopelessness in a sample of depressedoutpatients. Methods: In this study, we examined potential demographic, diagnostic, and symptom predictors of hopelessnessin a sample of 138 medication-free outpatients (73 women and 65 men) with a primary diagnosis of major depression. Thesignificance of predictors was evaluated in both simple and multiple regression analyses. Results: Consistent with previousstudies, we found no significant associations between demographic and diagnostic variables and greater hopelessness.Hopelessness was significantly associated with greater depression severity, poor problem solving abilities as assessed by theProblem Solving Inventory, and each of two measures of dysfunctional cognitions (the Dysfunctional Attitudes Scale and theCognitions Questionnaire). In a stepwise multiple regression equation, however, only dysfunctional cognitions and poorproblem solving offered non-redundant prediction of hopelessness scores, and accounted for 20% of the variance in thesescores. Limitations: This study is based on depressed patients entering into an outpatient treatment protocol. All analyseswere correlational in nature, and no causal links can be concluded. Conclusions: Our findings, identifying clinical correlatesof hopelessness, provide clinicians with potential additional targets for assessment and treatment of suicidal risk. Inparticular, clinical attention to dysfunctional attitudes and problem solving skills may be important for further reduction ofhopelessness and perhaps suicidal risk. 1999 Elsevier Science B.V. All rights reserved.

Keywords: Hopelessness; Depression; Dysfunctional attitudes; Problem solving; Suicidal risk

1. Introduction

*Corresponding author. Tel.: 1 1-617-724-5600; fax: 1 1-617- A negative view of the future, or hopelessness, has726-7541.

been identified as a regular cognitive correlate ofE-mail address: anierenberg@partners.org (A.A. Nierenberg)q depression (Beck et al., 1979; Beck et al., 1985), andThis study was supported in part by a grant from Eli Lilly and

Company. perhaps as a marker of select depression subtypes

0165-0327/99/$ – see front matter 1999 Elsevier Science B.V. All rights reserved.PI I : S0165-0327( 98 )00123-2

46 B. Cannon et al. / Journal of Affective Disorders 55 (1999) 45 –49

(Abramson et al., 1989). In clinical samples, hopel- with low hopelessness were more likely to haveessness has been shown to be a strong predictor of young children and to have had more positive lifesuicidal behavior. In a study of 68 consecutive events in the recent past (Greene, 1989).suicide attempters admitted to a general hospital, In an another examination of inpatients, suicidalhopelessness was found to correlate with suicidal intent appeared to covary with interpersonal problemintent more strongly than severity of depression solving abilities and mood during initial treatment(Minkoff et al., 1973). In a study of 1794 outpati- (Schotte et al., 1990). Problem solving abilities areents, hopelessness was 1.3 times as important as also linked to levels of dysfunctional attitudes, anddepression severity for estimating suicidal ideation both appear to covary with the level of mood(Beck et al., 1993). In another study of 207 patients disturbance (Miranda et al., 1990; Otto et al., 1997).admitted to the psychiatric inpatient wards of a Despite their association with current mood state,general hospital with suicidal ideation, neither the problem solving abilities and dysfunctional attitudesBeck Depression Inventory nor the Beck Scale for may have an independent association with levels ofSuicidal Ideation predicted the suicide of 14 patients hopelessness, as suggested by cognitive–behavioralduring the study’s ten-year follow-up period. Only accounts (Abramson et al., 1989; Beck et al., 1979).high scores on the Beck Hopelessness Scale differen- Identification of the correlates of hopelessness istiated patients who ultimately committed suicide important because of its potential value in helping(Beck et al., 1985). clinicians identify patients with higher suicidal risk.

Despite its apparent importance in predicting In addition, identification of the correlates of hopel-suicide, relatively few studies have examined the essness may help clinicians expand their targets forcorrelates of hopelessness. In a study of 80 depressed intervention. This is particularly important wheninpatients, Whisman et al. (1995) found no signifi- severity of depression, often the primary symptom ofcant differences between patients with high and low attention, is not well correlated with hopelessness. Inhopelessness on demographic characteristics, single the present study, we examine clinical correlates ofversus recurrent depression, number of prior epi- hopelessness in a large sample of outpatients withsodes of depression, and presence of co-occurring major depression. In particular, we examined levelsmelancholia, dysthymia, alcohol abuse, and drug of dysfunctional attitudes and poor problem solvingabuse. Increased hopelessness was associated with that may well contribute to feelings of hopelessness.both increased suicidal ideation and with increased We also examined correlations between hopelessnesssocial dysfunction, when controlled for severity of and a number of demographic and diagnostic vari-depression (Whisman et al., 1995). These findings ables. In the past, these variables failed to predictare consistent with an earlier report for a series of 88 levels of hopelessness (Whisman et al., 1995;outpatients with major depression; hopelessness did Nekanda-Trepka et al., 1983; Greene, 1989), butnot correlate with age, level of education, social could emerge as important under conditions of theclass, duration of illness, severity of symptoms, level greater statistical power and confidence in findingsof depression, or gender (Nekanda-Trepka et al., afforded by an adequate sample size.1983). Similar findings were evident in two studiesusing relatively small sample sizes. In a heteroge-neous sample of 20 patients diagnosed with depres-sion, hopelessness did not correlate with age, social 2. Methodsclass, number of prior hospitalizations, or length oftreatment. Hopelessness did correlate with a number Participants were 138 consecutive outpatients (65of problem life areas, pessimism, expectation of males, 73 females; mean age 396standardunpleasant consequences, and low confidence in the deviation 5 10.4) with major depression who wereability to bring about improvement (Nekanda-Trepka evaluated in a drug-free state with the Structuredet al., 1983). In another study that compared ten Clinical Interview for DSM-III-R (SCID-P) (Spitzerdepressed inpatients with low hopelessness and ten et al., 1989) prior to entry into a open treatment trial.depressed inpatients with high hopelessness, patients Exclusion criteria were a history of organic mental

B. Cannon et al. / Journal of Affective Disorders 55 (1999) 45 –49 47

disorders, history of seizure disorder, serious or Higher scores indicate more dysfunctional problemunstable medical illness, substance use disorders solving behavior. PSI scores were found to be(including alcohol) active within the last 12 months, significantly correlated with ratings of problem solv-serious suicidal risk, pregnancy, lactation, schizo- ing, and unrelated to perceived social desirability.phrenia, delusional disorder, psychotic disorders not Test–retest reliability for this measure has beenelsewhere classified, mood congruent or mood incon- found to be 0.89 for the total scale score (Heppnergruent psychosis, bipolar disorder, significant antiso- and Peterson, 1982).cial personality disorder, history of multiple adverse Depression severity was assessed by cliniciansdrug reactions or allergy to study drugs, non-re- using the 17-item Hamilton Rating Scale for Depres-sponse to an adequate course of antidepressant sion (Hamilton, 1960).treatment, concomitant use of birth control pills ornon-study psychotropic drugs, and clinical or labora-tory evidence of hypothyroidism. No patients wereexcluded due to suicidal risk. 3. Results

At baseline, patients completed a number of self-report questionnaires. Hopelessness was measured Demographic information and means and standardwith the Beck Hopelessness Scale (Beck and Weis- deviations for each of the severity measures aresman, 1974), a 20-item, self-rated questionnaire with presented in Table 1. The mean (6standard devia-a true–false response format which measures the tion) Beck Hopelessness Scale score was 11.665.1.extent of pessimistic outlook. Higher scores indicate The Beck Hopelessness Scale scores can be inter-more hopelessness. Factor analysis of hopelessness preted as follows: Asymptomatic (0 to 3), mild (4 toscale items previously has revealed affective, moti- 8), moderate (9 to 14) and severe (14 to 20) (Beckvational, and cognitive components. Test–retest re- and Steer, 1988). The mean age of onset of depres-liability for this measure was found to be 0.93 (Beck sion was 25.4612.5. Sixty-five patients (47%) metand Weissman, 1974). criteria for past or present anxiety disorder diag-

Dysfunctional attitudes were assessed with two noses, 43 (31%) met criteria for past alcohol abusemeasures: the Dysfunctional Attitude Scale (DAS) or dependence, and 33 (24%) met criteria for past(Weissman, 1979) and the Cognitions Questionnaire non-alcohol substance abuse or dependence. The(CQ) (Fennell and Campbell, 1984). The DAS is a mean Hamilton Rating Scale for Depression scoreself-rated questionnaire which measures cognitive was 19.563.4, reflecting a level of severity sufficientdistortions, fixed negative values and perfectionist to enter a clinical trial. Previous episodes of depres-attitudes associated with depression. In particular, it sion were common, with a mean of 3.465.4 episodestargets beliefs about specific contingencies determin- per patient (median episodes 5 2), according to dataing behavior and self-worth. The DAS consists of provided by a subset of 201 patients. None of the40-items which are rated on a scale from 1 to 7, patients reported non-response to a previous,indicating the degree of belief in the dysfunctional adequate trial of pharmacotherapy.attitude. Test–retest reliability for this measure hasbeen found to be 0.84 (Dobson and Breiter, 1983).

Table 1The Cognitions Questionnaire (Fennell and Camp- Means and standard deviations for continuous predictor variablesbell, 1984) is an additional measure of dysfunctional for the sample of depressed outpatientsattitudes which assesses dimensions of negative Variable Mean standard deviationthinking in relation to a number of hypothetical

Age 39.7 10.4events.HAM-D-17 19.5 3.4

Problem solving abilities were assessed with the Age of onset major depression 25.4 12.5Problem Solving Inventory (PSI) (Heppner and Number of previous episodes 3.4 5.4

Cognitions Questionnaire 27.6 10.7Peterson, 1982). The PSI is a 35-item questionnaireDysfunctional Attitudes Scale 148 34.8with each item scoring between and 1 and 6,Problem Solving Inventory 114 24.6assessing problem solving behavior and attitudes.

48 B. Cannon et al. / Journal of Affective Disorders 55 (1999) 45 –49

The significance of these variables for predicting provided non-redundant prediction of hopelessnesslevels of hopelessness were evaluated using simple scores, although the DAS could be substituted for theregression analysis. Continuous predictor variables CQ for nearly identical results. These variablesincluded patient age, age of onset of the first episode accounted for one-fifth of the variance in hopeless-of major depression, number of episodes of depres- ness scores.sion, and HRDS, DAS, CQ, and PSI scores. Cate- Our study does not allow identification of a causalgorical analysis included patient sex, a diagnosis of link between dysfunctional attitudes, poor problempast alcohol abuse or dependence or past non-alcohol solving and hopelessness. These variables have beensubstance abuse or dependence, and past or present hypothesized to place patients at risk for depression,anxiety disorders. To examine redundancy among demoralization, and hopelessness (Abramson et al.,predictors and help control the inflation of alpha in 1989; Beck et al., 1993; Nezu, 1987; Olinger et al.,these multiple comparisons, all variables were sub- 1987), and as such it is tempting to assume thatsequently examined in a stepwise multiple regression hopelessness is the result of the negative and dis-analysis. couraging world view fostered by maladaptive at-

The simple regression analysis revealed four sig- titudes and poor problem solving skills. However, wenificant predictors of hopelessness scores: the CQ cannot rule out the alternative explanation that(r 5 0.42, p , 0.0001), DAS (r 5 0.35, p , 0.0001), maladaptive attitudes and poor problem solving skillsthe HRSD (r 5 0.21; p , 0.01), and the PSI (r 5 are a consequence of a more general sense of despair0.38; p , 0.0001). In all cases, the relationship that inhibits adaptive functioning (cf., Otto et al.,between each predictor and the hopelessness scores 1997). In addition, our findings are necessarilywas positive; greater dysfunctional cognitions and specific to the type of patients entered in this studyattitudes, depression severity, and dysfunctional and encountered in outpatient practice; i.e., patientsproblem solving were each associated with greater who do not face current, serious suicidal risk, andhopelessness. However, in the stepwise, multiple should be immediately hospitalized.regression equation only the CQ and the PSI offered Identifying correlates of hopelessness providesnon-redundant predictability of hopelessness scores clinicians with additional targets for assessment and(R 5 0.45, df 5 2, 135; p , .0001), and accounted treatment of suicidal risk. Indeed, there is initialfor 20% of the variance in these scores, although the evidence that cognitive therapy, which attends direct-DAS could be substituted for the CQ for nearly ly to dysfunctional attitudes and problem solvingidentical results. skills in depression, is more effective in decreasing

hopelessness than pharmacologic interventions (Rushet al., 1982). These findings encourage clinicians to

4. Discussion attend to more than the severity of depression injudging the adequacy of their interventions and

Consistent with previous studies (Whisman et al., assessing suicidal risk. Attention to these additional1995; Nekanda-Trepka et al., 1983; Greene, 1989), correlates of hopelessness (problem solving skillswe found no relationship between demographic and and dysfunctional attitudes) may provide the clini-diagnostic variables and hopelessness in a large cian with the greatest opportunity to decrease suici-sample of outpatients with major depression. Al- dal risk.though severity of depression was significantly asso-ciated with hopelessness, its predictive significancewas redundant with other, more powerful predictors.

ReferencesTwo measures of dysfunctional cognitions (the DASand CQ) and the PSI were each significantly and

Abramson, L.Y., Metalsky, G.I., Alloy, L.B., 1989. Hopelessnesspositively associated with levels of hopelessness. Indepression: a theory-based subtype of depression. Psychol.

a stepwise multiple regression equation, only Rev. 96, 358–372.dysfunctional cognitions as measured by the CQ and Beck A.T., Rush A.J., Shaw B.F., Emery G., 1979. Cognitiveimpaired problem solving (assessed with the PSI) Therapy of Depression. Guilford Press, New York.

B. Cannon et al. / Journal of Affective Disorders 55 (1999) 45 –49 49

Beck A.T., Steer R.A., 1988. Manual for the Beck Hope- Nekanda-Trepka, C.J.S., Bishop, S., Blackburn, I.M., 1983.lessness Scale. Psychological Corporation, San Antonio, Texas. Hopelessness and depression. Brit. J. Clin. Psychol. 22, 49–60.

Beck, A.T., Steer, R.A., Beck, J.S., Newman, C.F., 1993. Hope- Nezu, A.M., 1987. A problem solving formulation of depression:lessness, depression, suicidal ideation, and clinical diagnosis of A literature review and proposal of a pluralistic model. Clin.depression. Suicide and Life-Threatening Behavior 23, 139– Psychol. Rev. 7, 121–144.145. Olinger, L.J., Kuiper, N.A., Shaw, B.F., 1987. Dysfunctional

Beck, A.T., Steer, R.A., Kovacs, M., Garrison, B., 1985. Hopel- attitudes and stressful life events: An interactive model ofessness and eventual suicide: A 10-year prospective study of depression. Cogn. Ther. Res. 11, 25–40.patients hospitalized with suicidal ideation. Am. J. Psychiatry Otto, M.W., Fava, M., Penava, S.A., Bless, E., Muller, R.T.,142, 559–563. Rosenbaum, J.F., 1997. Life event and cognitive predictors of

Beck, A.T., Weissman, A., 1974. The measurement of pessimism: perceived stress before and after treatment for major depres-The hopelessness scale. J. Consult. Clin. Psychol. 42, 861–865. sion. Cog. Ther. Res. 21, 409–420.

Dobson, K.S., Breiter, H.J., 1983. Cognitive assessment of Rush, A.J., Beck, A.T., Kovacs, M., Weissenburger, J., Hollon,depression: Reliability and validity of three measures. J. S.D., 1982. Comparison of the effects of cognitive therapy andAbnorm. Psychol. 92, 107–109. pharmacotherapy on hopelessness and self-concept. Am. J.

Fennell, M.J.V., Campbell, E.A., 1984. The cognitions ques- Psychiatry 139, 862–866.tionnaire: Specific thinking errors in depression. Brit. J. Clin. Schotte, D.E., Geds, J., Payvar, S., 1990. Problem-solving deficitsPsychol. 23, 81–92. in suicidal patients: trait vulnerability or state phenomenon? J.

Greene, S.M., 1989. The relationship between depression and Consult. Clin. Psychol. 58, 562–564.hopelessness: implications for current theories of depression. Spitzer, R.L., Williams, J.B.W., Gibbon, M.B., First, M. (Eds.),Brit. J. Psychiatry 154, 650–659. 1989. Structured Clinical Interview for DSM-III-R-Patient

Hamilton, M., 1960. A rating scale for depression. J. Neurol. Version (SCID-P, 9 /1 /1989 version). New York BiometricsNeurosurg. Psychiatry 23, 56–62. Research Department, New York State Psychiatric Institute.

Heppner, P.P., Peterson, C.H., 1982. The development and impli- Weissman, A., 1979. The Dysfunctional Attitude Scale: A valida-cations of a personal problem solving inventory. J. Counsel. tion study. Dissert. Abs. International 40 (B), 1389–1390.Psychol. 29, 66–75. Whisman, M.A., Miller, I.W., Norman, W.H., Keitner, G.I., 1995.

Minkoff, K., Bergman, E., Beck, A.T., Newman, C.F., 1973. Hopelessness depression in depressed inpatients: symp-Hopelessness, depression, and attempted suicide. Am. J. Psy- tomatology, patient characteristics, and outcome. Cogn. Ther.chiatry 130, 455–459. Res. 19, 377–398.

Miranda, J., Persons, J.B., Byers, C.N., 1990. Endorsement ofdysfunctional beliefs depends on current mood state. J. Ab-norm. Psychol. 99, 237–241.