DOSS: Delirium Observation Screening Scale

This original development paper tested a 25-item DOS scale—not the later 13-item clinical form—in 82 geriatric-medicine and 92 hip-fracture patients. It reported high internal consistency and good predictive validity against DSM-IV geriatrician diagnosis, but only 22 participants developed delirium across the two cohorts.

This academic-hospital study compared the DOSS with the DRS-R-98 in 101 assessments of 54 patients aged older than 64 years. It reported sensitivity of 90% and specificity of 91%, although the small cohort and use of repeated assessments limit patient-level generalisability.

This study compared ward-nurse DOSS screening with daily CAM assessment in 138 cognitively impaired inpatients aged 70 years or older, 44 of whom developed delirium. Sensitivity was 56% and specificity 92%; detection fell to 38% for hypoactive and 60% for mixed delirium, supporting addition of a formal attention test.

This advanced-cancer study compared DOSS scores with the DRS-R-98 in 95 screen-negative and 98 screen-positive patients at a threshold of 3 or more. It reported sensitivity above 99.9% and specificity 99.5%, but the deliberately selected positive and negative groups and specialist setting limit transportability to unselected routine populations.

Trained nurses assessed 698 patients aged 65 years or older on three general medical wards, with neurologists applying DSM-5 criteria. DOSS sensitivity was 94%, specificity 86%, positive predictive value 40% and negative predictive value 99% at 9% delirium prevalence; the study was confined to one hospital.

This DOSS-specific diagnostic meta-analysis included eight studies and used a hierarchical regression model. Pooled sensitivity was 90% (95% CI 76%–97%) and specificity 92% (95% CI 88%–94%); the wide sensitivity interval and between-study differences matter when applying the average to a particular service.

This systematic review included 22 research studies and four audit reports of delirium tools used at scale in routine general-hospital care. DOSS inpatient positive-score rates ranged from 6% to 42% and overall completion from 19% to 100%; most studies had moderate-to-high risk of bias, showing that validation performance does not guarantee reliable implementation.

This six-site study compared DOSS with CAM ratings made by three trained nurse researchers in 338 long-term-care residents. At the threshold of 3 or more, the released dataset yields sensitivity of 97.1% and specificity 95.4%; trained research conditions and use of CAM as comparator differ from routine staff use and expert diagnostic assessment.

This scoping review included 58 nursing-home and long-term-care studies, identifying 25 detection tools but only 14 with setting-specific validity or reliability testing. DOSS had the highest reported diagnostic accuracy, while the review highlighted poor standardisation, limited training and overlap with dementia; it did not produce pooled sensitivity or specificity.