One of the problems in my field is outcome scales: they are all proxy to what really interests us. Positively, we want people to get better (be well) or at least have recovery (Live well with disability). Negatively, we want to see less death, less disability, less disadvantage.
But we use scales. And even on the most important outcomes, such as suicide, there is considerable disagreement.
To study the disagreement between self-reported suicidal ideation (SR-SI) and clinician-ascertained suicidal ideation (CA-SI) and its correlation with depression and anxiety severity in patients with major depressive disorder (MDD) or bipolar disorder (BPD).
Methods
Routine clinical outpatients were diagnosed with the MINI-STEP-BD version. SR-SI was extracted from the 16 Item Quick Inventory of Depression Symptomatology Self-Report (QIDS-SR-16) item 12. CA-SI was extracted from a modified Suicide Assessment module of the MINI. Depression and anxiety severity were measured with the QIDS-SR-16 and Zung Self-Rating Anxiety Scale. Chi-square, Fisher exact, and bivariate linear logistic regression were used for analyses.
Results
Of 103 patients with MDD, 5.8% endorsed any CA-SI and 22.4% endorsed any SR-SI. Of the 147 patients with BPD, 18.4% endorsed any CA-SI and 35.9% endorsed any SR-SI. The agreement between any SR-SI and any CA-SI was 83.5% for MDD and 83.1% for BPD, with weighted Kappa of 0.30 and 0.43, respectively. QIDS-SR-16 score, female gender, and ?4 year college education were associated with increased risk for disagreement, 15.44 ± 4.52 versus 18.39 ± 3.49 points (p = 0.0026), 67% versus 46% (p = 0.0783), and 61% versus 29% (p = 0.0096). The disagreement was positively correlated to depression severity in both MDD and BPD with a correlation coefficient R2 = 0.40 and 0.79, respectively, but was only positively correlated to anxiety severity in BPD with a R2 = 0.46.
Conclusion
Self-reported questionnaire was more likely to reveal higher frequency and severity of SI than clinician-ascertained, suggesting that a combination of self-reported and clinical-ascertained suicidal risk assessment with measuring depression and anxiety severity may be necessary for suicide prevention.
What should we make of this? As a clinician, do I mistrust my judgement or that of the patient? In the final analysis, it is the patient that acts: the number of times that the notes state clinical improvement and within a week we are having another enquiry is high. But if you reverse it and consider the patient’s thoughts you have to reconcile the death rate — which even for patients of the mental system is around one in a thousand for suicide — with one in five reporting ideation.
What we are dealing with here is a category error. The thought is not the action. The scale is not the act. The person is not the file. And life is only neat and definite in the minds of judges.