Engaging families to detect ICU delirium

Fair diagnostic accuracy was achieved when family members used delirium tools for daily assessments.


Detecting delirium is notoriously difficult in the ICU. One challenge is that hospital clinicians often don't know how patients usually behave. Recognizing that their family members do, researchers at the Foothills Medical Centre ICU recently involved them in the testing of layperson use of delirium diagnosis tools.

“Our group in Calgary has had a long tradition of engaging patients and families as partners in our research, so I thought that family partnership with delirium would be an interesting thing to study,” said Kirsten M. Fiest, PhD, an assistant professor of critical care medicine and director of research and innovation in the department of critical care medicine at the University of Calgary Cumming School of Medicine.

How it works

Between December 2017 and March 2019, the researchers recruited 147 ICU patients (mean age, 56.1 years; 61% men) and family members (48.3% spouses). For a maximum of five days, families completed daily assessments of the patients using the Family Confusion Assessment Method (FAM-CAM) and Sour Seven tools. During the same period, a board-certified neuropsychiatrist and team of ICU research nurses conducted reference standard assessments of the patients using the Intensive Care Delirium Screening Checklist and Confusion Assessment Method for ICU.


Delirium was present in 37.0% of 508 reference-standard assessments, according to results published in the July Critical Care Medicine. The sensitivity of the FAM-CAM, compared with the reference standard, was 54.1%, with a specificity of 76.8%. On the Sour Seven, the sensitivity and specificity for possible delirium (cutpoint of 4) were 72.9% and 68.8% and for delirium (cutpoint of 9) were 51.1% and 82.8%, respectively, compared with the reference standard. The tools took four to seven minutes to complete, compared to 10 to 15 minutes for the clinician-administered tools.

While the family-administered tools had fair diagnostic accuracy, the results were lower than those of the clinician-administered assessments. “However, I don't think they're so low to the point of not being a potential option,” said Dr. Fiest. In addition, a follow-up study found that the delirium detection helped family members feel engaged as part of the care team, she reported. “So even if the tool maybe isn't as effective as we'd hoped, if it gives the family this sense of purpose and makes them feel better and feel involved, I think that was also a positive.”


The biggest challenge was that family was not necessarily present at the bedside every day, which meant some serial assessments over the five-day period were incomplete, said Dr. Fiest. “One of the reasons that we decided that was OK was that we had codesigned our study with former ICU patients and their family members, and they said flexibility was really key,” she said. “At such a stressful moment for them, the pressure of having to do something every day isn't always going to work out.”

Another challenge was that families may not have understood what they were looking for. While the delirium-detection tools were initially developed to be used by laypeople, they were developed for use in general medical ward patients or older adult outpatients, Dr. Fiest noted. “So perhaps the ICU environment adds a level of complexity, especially if a patient is intubated, for example, or might be more sedated than normal.”

Lessons learned

The team initially conducted a pilot study to determine the feasibility and acceptability of the project and found recruitment a challenge. “People were just not willing to participate,” said Dr. Fiest. So the team hired a family member of a former ICU patient as part of the study team, and she took over the recruitment and leveraged her rapport with other family members.

The involvement of patients and families in this kind of research is key, according to Dr. Fiest. “Not many centers do this, and those that do are probably ahead of the curve, but it totally transformed our ability to recruit. . . . We learned a lesson from the pilot study, which is why pilot studies are so important, because otherwise we would have potentially had a really atrocious response rate,” she said. “And partnering with someone with the lived experience that we were trying to understand better was also really helpful.”

Next steps

The next step is to combine delirium detection into a bundle based on other research on delirium prevention and management. This includes a validated educational module to train family members on what delirium is and how to detect it, as well as to coach them on prevention and management strategies, said Dr. Fiest. “Distinguishing delirium from dementia, for example, was something that family members struggled with,” she said. “Something that would be more helpful than just providing the family members with the questionnaire is to provide them some education first.”