JHQ 300: Efficacy of the Crisis Risk Triage Scale in Inpatient Units within the United States-JHQ 300 Article

JHQ 300 Article

Pdf Summary

This retrospective study evaluated the Crisis Triage Rating Scale (CTRS) as a screening and risk-stratification tool for inpatient violence risk in a 12-hospital integrated health system in the southeastern United States. The project was motivated by rising behavioral safety events and limitations of the previously used SAD PERSONS tool, which is focused on suicide risk and is not designed to assess risk of harm to others.<br /><br />Electronic health records from 112,708 adult inpatient visits (Jan 2019–Dec 2020) were reviewed; after exclusions for setting, missing/invalid documentation, and timing (CTRS recorded after an event), 87,566 visits with valid CTRS assessments were analyzed. Behavioral safety events were identified via the safety event reporting system. Because documented self-harm was extremely rare (7 events), analyses focused on harm to others (482 events), including combative behavior, verbal threats, assaults to staff, and security assists.<br /><br />The CTRS “Intent to Harm Self or Others” triage question and CTRS-derived risk levels (high 3–8, moderate 9–12, low 13–15) significantly predicted harm-to-others events. Logistic regression models showed increased odds of harm for patients scoring at-risk compared with those scored no-risk, with the highest-risk category showing the strongest association (unadjusted OR 13.83; adjusted OR 12.28). However, because harm events were infrequent, predictive performance showed very low sensitivity (0.029–0.171) and low positive predictive value (0.025–0.070), but very high specificity (0.931–0.999) and negative predictive value (0.991–0.995).<br /><br />The authors conclude that CTRS performs as expected in low-incidence environments: a positive score should be interpreted cautiously when allocating scarce safety resources, while a negative score is useful for ruling out risk. Practice variation in completing CTRS across the system suggests that improved adherence and documentation could strengthen risk stratification. Continuous evaluation and subgroup analyses are recommended.

Keywords

Crisis Triage Rating Scale (CTRS)

inpatient violence risk screening

harm-to-others prediction

behavioral safety events

risk stratification (high/moderate/low)

logistic regression odds ratios

low-incidence event prediction

sensitivity specificity PPV NPV

electronic health record retrospective study

SAD PERSONS tool limitations