How different ways of assessing suicidal ideation influence its prevalence, correlates and predictive validity among patients with major depressive disorder (MDD) remains unclear.

Within the Vantaa Primary Care Depression Study (PC-VDS, 91 patients) and the Vantaa Depression Study (VDS, 153 psychiatric out-and 41 inpatients), suicidal ideation was assessed with the Scale for Suicidal Ideation (SSI), Hamilton Depression Scale (HAM-D) item 3 and Beck Depression Inventory (BDI) item 9, and by asking whether patients had seriously considered suicide during the episode. The positive and negative predictive values (PPV, NPV) for suicide attempts during a six-month follow-up were investigated.

Depending on the setting, 56–88% of patients had suicidal ideation in some of the assessments, but only 8–44% in all of them. Agreement ranged from negligible to moderate (kappa 0.06–0.64), being lowest among primary care patients. The correlates of suicidal ideation overlapped. No assessment had optimal sensitivity, specificity, PPV and NPV. Nevertheless, PPVs ranged up to 43%.

Which MDD patient is classified as having suicidal ideation depends strongly on the method of assessment, with the greatest variation likely in primary care. Differences in assessments may cause inconsistency in risk factors. Predicting suicide attempts is difficult, but not futile.