Predictors of Persistent Thoughts of Suicide among Psychiatric Inpatients

Thomaskutty Idiculla, PhD

McLean Hospital – Faculty

Scientific Abstract

Background: Understanding the predictors of mental health disorders is particularly important when assessing patients with thoughts of self-harm and/or suicide. In 2019, 3.5 million Americans reported plans for committing suicide, and 1.4 million attempted suicide. 22% of suicides occurred within the first year of living with mental illness. In fact, 90% of people who die by suicide have an underlying and potentially treatable psychiatric disorder. BASIS-24 is a comprehensive 24-item self-reported assessment tool that can identify patient symptoms and functioning in specific areas of difficulty across all diagnostic groups, including thoughts of suicide or self-harm. The purpose of this study is to examine predicting factors for individuals who will experience persistent thoughts of self-harm and suicide.  

Method: The present study consists of 38,502 psychiatric inpatients who completed the BASIS-24 survey at admission and/or discharge from 2011 to 2021. Patients completed BASIS-24 at admission and discharge and had a valid ICD-10 discharge diagnosis. The sample consisted of 53.0% female patients, 87.2% were white, and the mean age of patients was 37.73 years old (SD= 15.21). Of these, 36.4% had a primary diagnosis of depression; 23.6% had alcohol/drug use disorders, 19.0% had bipolar disorder, and 9.6% had schizophrenia or schizoaffective disorder. At the time of admission, 60.4% of the total sample had thoughts of self-harm and/or suicide. At the time of discharge, 27.7% of the total sample had thoughts of self-harm and/or suicide.  BASIS-24 consists of two items related to thoughts of suicide and self-harm in the past week. One item is as follows: “How much of the time did you think about ending your life?” Patients responded using a 5-point Likert-type scale with responses ranging from “None of the Time” at 0, and “All of the Time” at 4. The second item is as follows: “How often did you think about hurting yourself?” Patients responded using a 5-point Likert-type scale with responses ranging from “Never” at 0 to “Always” at 4. For this study, the self-harm item was grouped by “no self-harm” and “serious self-harm”, such that patients who responded with 0 were put in the “no self-harm” group, and patients who responded with either 3 or 4 were put in the “serious self-harm” group. Logistic regression model was used to examine the correlates on the risk of self-harm at admission. The model was adjusted for age, gender, length of stay, race, marital status, employment status, primary diagnosis, disability, social support, and BASIS-24 subscales.

Results: The study indicates that BASIS-24 subscales and certain patient characteristics predict thoughts of self-harm or suicide. Logistic regression analysis showed that significant factors for persistent suicidal thoughts for intake sample (p<0.01) were BASIS-24 depression score, interpersonal relationships, emotional lability at admission; multiple comorbidities, longer length of stay, being female, being a young adult, never married, and lack of family or social support.  This study shows that BASIS-24 can be used for assessing suicide risk and predicting suicide risk with other patient characteristics. Future studies will focus on the predictability of suicide using BASIS-24 at discharge.


Live Zoom Session – March 9th

research Areas


Thomas Idiculla, PhD, Jason Berkowitz, BA, Caroline Howley, BS

Principal Investigator

Thomas Idiculla, PhD

Affiliated Website