Table 4.

Recommendations for Confusion Assessment Method training and oversight for multisite studies^a

Initial training and standardisation
Didactic overview	• Classroom: guided review of the CAM Training Manual
	• Interactive review and scoring of training videos
Individual practice sessions	• Paired practice interviews with an experienced delirium assessor
	• Mimic 2 patients with and 2 patients without delirium
Pilot interviews with patients	• Experienced delirium assessor observes new research assistant interviewing patients and gives feedback
	• Interview 2 delirium and 2 non-delirium patients
Inter-rater reliability assessments (baseline standardisation)	• Pairs of interviewers observe same patients and score CAM independently
	• After interview, compare and discuss ratings
	• Continue until 100% agreement achieved
	• Minimum of 2 delirium and 5 non-delirium patients
	• Early pairs should include experienced delirium assessor
CAM-only training	• Intended to score cases where patients are poorly responsive or interviews are incomplete
	• Training to code CAM features based on bedside observations. If patients unresponsive, may only be able to code altered level of consciousness
Ongoing monitoring and performance checks
Coding sessions	• Regularly scheduled meetings to discuss any questions on coding the CAM features
	• Involve project directors and key staff from each study site, and include at least one delirium expert clinician
	• Minimum of 2 times per month (ideally weekly) throughout the study
	• Use sessions as opportunity for retraining
Ongoing inter-rater reliability Assessments (performance checks)	• Local: all staff undergo paired ratings with experienced delirium assessor at the site; ratings compared and discussed. Recommend: every 6 months throughout study
	• Cross-site: One gold-standard expert rater performs spot checks at all study sites with inter-rater assessments at least once per year. Alternative approaches may utilise videoconferencing or face-time for inter-rater assessments across sites.
New staff training	• Complete all steps of initial training when any new staff member joins the study to maintain high quality ratings
	• Verify inter-rater reliability with existing staff

^aNote: All steps should be overseen by the central coordinating centre, and one fully-trained, experienced delirium assessor (principal investigator, project director, or experienced research staff member) is required at each site to provide ongoing monitoring and training locally. For optimal training, all raters should be trained by an experienced CAM rater