A new study suggests that suicidality assessment during ECT may depend partly on which depression questionnaire a patient completes. Among older adults with treatment-resistant depression, two widely used measures did not always identify suicidal thoughts during the same treatment sessions.
The findings highlight a clinically important distinction: agreement between screening tools does not guarantee that either tool captures every patient experience requiring further evaluation.
Why Suicidality Assessment During ECT Demands Precision
Electroconvulsive therapy is an established treatment for severe depression, including cases that have not improved with medication or psychotherapy. It may be considered when symptoms are profound, functioning is severely impaired, or a rapid clinical response is needed.
Because suicidal thinking can change during an acute ECT course, clinicians often use patient-reported depression scales to track symptoms over time. Two common instruments are the Patient Health Questionnaire-9 and Beck Depression Inventory-II.
Each contains one item related to suicidal thinking or self-harm. However, the items use different wording, and patients may not interpret them in the same way.
Two Familiar Scales Produced Different Signals
Researchers analyzed data from 97 patients aged 60 or older who had treatment-resistant depression and received an acute course of ECT. Every included patient completed both the PHQ-9 and BDI-II and had consented to the use of their information for research.
A response greater than zero on the relevant item of either questionnaire was classified as indicating suicidality. Across 1,251 ECT sessions with complete data, the investigators compared whether the two scales agreed on the presence or absence of these symptoms.
This repeated assessment model offered a session-by-session view rather than relying on a single measurement taken before or after treatment.
Disagreement Appeared Across More Than 100 Sessions
Seventy-one of the 97 patients, or 73 percent, endorsed suicidality on at least one measure at some point during treatment. Overall, suicidal thoughts or self-harm content appeared in 430 sessions, representing 34.4 percent of the observations.
Both measures were positive during 305 sessions. However, the PHQ-9 alone identified suicidality during 81 sessions, while the BDI-II alone identified it during 44.
The association between the two measures was statistically significant. Still, the clinically notable finding was their incomplete overlap. Nearly one-third of positive sessions were identified by only one questionnaire.
Greater Detection Does Not Automatically Mean Greater Accuracy
The PHQ-9 captured almost twice as many unique positive sessions as the BDI-II. This pattern may indicate greater sensitivity in this particular population, but the study does not establish that every additional PHQ-9 response represented suicidal intent.
The PHQ-9 item combines thoughts of being better off dead with thoughts of self-harm. A positive response can therefore reflect several experiences with different levels of immediacy and risk. Previous research has also found that PHQ-9 item 9 predicts elevated risk while remaining insufficient as a standalone suicide risk assessment.
The results should consequently be interpreted as evidence of measurement discordance, not proof that one questionnaire can definitively classify suicide risk.
Wording May Shape What Patients Disclose
Small differences in language can change how a person understands a sensitive question. One item may feel more relevant to passive thoughts about death, while another may prompt disclosure of active self-harm thinking.
Age, cognitive symptoms, fatigue, treatment timing, and comfort with specific wording may also influence responses. During an ECT course, these factors can shift from one session to another.
That makes a questionnaire response a starting point for clinical conversation rather than a complete account of risk.
Repeated Measurement Reveals What One-Time Screening Can Miss
A key strength of this analysis is its use of data from more than 1,200 treatment sessions. The design shows that discrepancies can recur throughout care, even when the same patients complete both measures.
At the same time, repeated sessions came from only 97 individuals. Future studies could examine how within-patient changes, cognitive effects, item wording, and clinician-administered interviews influence agreement.
Clinical Monitoring May Need More Than One Signal
For ECT programs, the practical implication is not necessarily that every patient requires multiple lengthy questionnaires. Instead, teams may need a process for investigating positive responses, unexpected score changes, and concerns that arise outside formal scales.
Structured follow-up questions and direct clinical assessment remain essential. A negative response on one depression measure should not automatically override other evidence, including a conflicting questionnaire, behavioral changes, caregiver concerns, or the patient’s clinical history.
This study adds a focused warning to measurement-based care: standardized scales can improve consistency, but their results are shaped by the questions they ask. Better suicidality assessment during ECT may depend on combining repeated measurement with individualized clinical judgment.
Citations
Mandell A, et al. “S28. Discrepancies in Suicidality Assessments Among Geriatric TRD Patients Receiving ECT.” The American Journal of Geriatric Psychiatry, 2026. https://www.sciencedirect.com/science/article/abs/pii/S1064748126002903
Na PJ, et al. “The PHQ-9 Item 9 Based Screening for Suicide Risk: A Validation Study of the Patient Health Questionnaire.” Journal of Affective Disorders, 2018. https://pubmed.ncbi.nlm.nih.gov/29477096/
