Skip to main content
. Author manuscript; available in PMC: 2023 Mar 2.
Published in final edited form as: JAMA Psychiatry. 2016 Oct 1;73(10):1064–1071. doi: 10.1001/jamapsychiatry.2016.2172

Table 5.

Cox Proportional Hazards Regression Model of Suicide Risk Without (Model 1) and With (Model 2) Inclusion of Natural Language Processing of Narrative Discharge Summary

Variable Adjusted Hazard Ratio (95% CI)a
Model 1 Model 2
Male sex 2.80 (2.11–3.71) 2.57 (1.94–3.41)
White race 2.27 (1.56–3.30) 2.23 (1.53–3.24)
Age at admission, y 0.80 (0.65–0.97) 0.79 (0.65–0.96)
Age-adjusted Charlson Comorbidity Index 0.95 (0.74–1.21) 0.89 (0.69–1.15)
Primary psychiatric diagnosis at admission 1.63 (0.93–2.85) 1.39 (0.79–2.45)
Psychiatric visits in 12 mo before admission 1.18 (1.07–1.31) 1.17 (1.06–1.30)
Prior psychiatric visits (ever) 1.15 (1.01–1.31) 1.14 (1.01–1.30)
Outpatient visits (any) in 12 mo before admission 0.79 (0.67–0.94) 0.84 (0.71–1.00)
Emergency department visits in 12 mo before admission 2.37 (1.83–3.06) 2.19 (1.68–2.84)
Positive valence NA 0.70 (0.58–0.85)
Negative valence NA 1.15 (0.99–1.32)

Abbreviation: NA, not applicable.

a

Adjusted for all other terms in the model.