Table 5.

Longitudinal Studies Investigating the Relationship Between Insight and Depression

Study	Design	Participants	Measures	Results
Smith et al73	Assessment at discharge and after an average of 20.4 d for inpatients and 86.7 d for outpatients	33 patients with schizophrenia and schizoaffective disorders (Structured Clinical Interview for DSM-IV)	Insight: SUMD	Unawareness of current and past symptoms and misattribution of current symptoms were correlated with depression at time one (r = −.46–.58). After entering depression, disorganized symptoms, positive symptoms, and treatment unit in linear regression analyses, depression was the only significant predictor for unawareness and misattribution of current symptoms (R2 = 28–31%). Increase in depression by 2. Assessment was associated with more awareness
			Depression: BPRS
			Symptoms: SANS, SAPS
Crumlish et al106	Assessment at presentation to clinic, at 6 mo and 4 y	101 patients with first episode of schizophrenia or schizoaffective disorder	Insight: PANSS insight, IS.	The PANSS lack of insight items at presentation and at 6 mo did not predict depression at 4 y or suicide in the follow-up period. Using the Birchwood Scale, greater insight at 6 mo predicted higher depression at 4 y and suicide in the follow-up period.
			Depression: PANSS anxiety, depression, somatic concern, and guilt items
Drake et al107	Assessment within 14 d after admission, reassessments after 6 and 12 wk and 18 mo	257 patients with schizophrenia, schizoaffective, schizophreniform, delusional disorder, or psychosis not otherwise specified (DSM-IV)	Insight: IS, PANSS insight	The association between depression and insight at baseline was r = .43, and for insight and self-esteem r = .37. Cross-lagged associations were generally weak. The pattern of associations was similar at each stage: Better insight predicted more depression and was associated with lower self-esteem. Greater paranoia predicted greater depression. The link between insight and depression was not just mediated by low self-esteem or paranoia.
			Depression: PANSS anxiety, depression, avolition, and guilt
			Self-esteem: Rosenberg Self-Esteem Scale
Smith et al108	Assessment at discharge and follow-up at 6 mo)	50 patients with schizophrenia or schizoaffective disorder (SCID for DSM-IV)	Insight: SUMD	Depression at discharge was not associated with current awareness of positive or negative symptoms, but patients with depression showed more improvement of insight into past symptoms (ß = −.36 in regression model)
			Depression: BPRS
			Symptoms: BPRS
Iqbal et al109	Assessment during acute stage, upon recovery and 12 mo after recovery	105 patients with ICD-10 schizophrenia (of which 78% completed all assessments)	Insight: IS	Prior to postpsychotic, patients reported more negative appraisals of the disorder and lower self-esteem in comparison to those without postpsychotic depression but did not differ in insight. During depression, patients reported higher insight into disorder, increase in negative appraisals, and decrease in self-esteem.
			Attitudes: PBIQ
			Depression: BDI, CDSS, DEQ
			Self-esteem: Self-Esteem Scale
Caroll et al75	Assessment before discharge, random assignment to intervention or control group and reassessment after 12 wk	100 patients with DSM-III-R schizophrenia (with at least one previous psychotic episode in the past)	Insight: ITAQ	At baseline, insight correlated positively with depression (r = .23). 22% of the variance of baseline insight could be explained by gender, depression, positive and negative symptoms, age, and illness duration. In the final equation, only positive symptoms (ß = .43) and MADRS scores (ß = .26) made significant contributions to predicting insight.
			Symptoms: PANSS	In the regression analysis to predict change in insight by change in positive and negative symptoms and depression, only MADRS change contributed significantly (ß = .24)
			Depression: MADRS
Yen et al76	Assessment in out and inpatient settings, follow-up at 1 y.	74 patients with DSM-IV schizophrenia in remission	Insight: SAI	One patient had committed suicide by follow-up and 7 had planned suicide attempts. Insight was not significantly correlated with suicidal tendency during the 1-y follow-up.
			Symptoms: PANSS
			Suicide: VASA
Bourgeois et al110	Weekly assessment of suicide events in months 1–6, biweekly in months 6–24. Assessment of depression and awareness at weeks 24, 52, and 104.	980 patients with schizophrenia or schizoaffective disorder and high risk for suicide	Insight: Item 12 from Scale of Functioning	Results of Cox proportional hazards analysis for main effects of baseline variables demonstrated that greater awareness significantly predicted suicide events (hazards ratio = 1.17). However, the inclusion of depression at in the model baseline rendered awareness nonsignificant.
			Depression: CDSS
			Suicide: blinded judgments of suicide events (suicide attempt or hospitalization to prevent suicide)	Although baseline awareness was a risk factor, increases in awareness as a function of treatment were associated with a decreased risk of suicide events (hazards ratio = 0.75).

Note: DSM-IV, Diagnostic and Statistical Manual for Mental Disorders, Fourth Edition; ICD-10, International Classification of Diseases, 10th Revision; PBIQ, Personal Beliefs About Illness Questionnaire; BDI, Beck Depression Inventory; CDSS, Calgary Depression Scale for Schizophrenia; DEQ, Depressive Experiences Questionnaire; DSM-III-R, Diagnostic and Statistical Manual for Mental Disorders, Revised Third Edition; MADRS, Montgomery and Asberg Depression Rating; VASA, Violence and Suicide Assessment Scale.