The DEMS-DOSS study: validating a delirium monitoring tool in hospitalised older adults.

Objective: to evaluate the sensitivity, specificity and test-retest reliability of the Delirium Early Monitoring System-Delirium Observation Screening Scale (DEMS-DOSS).

Design: prospective diagnostic accuracy study of a convenience sample of admitted older adults with DEMS-DOSS and reference standard assessments.

Setting: 60-bed aged care precinct at a metropolitan hospital in Sydney, Australia.

Participants: 156 patients (aged ≥65 years old) were recruited to participate between April 2018 and March 2020. One hundred participants were included in the analysis.

Measurements: Participants were scored on the DEMS-DOSS. Trained senior aged care nurses conducted a standardised clinical interview based on the Diagnostic and Statistical Manual of Mental Disorder (DSM)-IV delirium criteria, within two hours of DEMS-DOSS completion. The senior aged care nurse undertaking the DSM-IV interview was blinded to the results of the DEMS-DOSS.

Results: Participants' mean age was 84 (SD ±7.3) years and 39% (n = 39) had a documented diagnosis of dementia. Delirium was detected in 38% (n = 38) according to the reference standard. The DEMS-DOSS had a sensitivity of 76.3% and a specificity of 75.8% for delirium. The area under the receiver operating characteristics curve for delirium was 0.76. The test-retest reliability of the DEMS-DOSS was found to be high (r = 0.915).

Conclusion: DEMS-DOSS is a sensitive and specific tool to assist with monitoring new onset and established delirium in hospitalised older adults. Further studies are required to evaluate the impact of the monitoring tool on health outcomes.