A brief intervention for preparing ICU families to be proxies: A phase I study

BACKGROUND: Family members of critically ill patients report high levels of conflict with clinicians, have poor understanding of prognosis, struggle to make decisions, and experience substantial symptoms of anxiety, depression, and post-traumatic stress regardless of patient survival status. Efficient interventions are needed to prepare these families to act as patient proxies. OBJECTIVES: To assess a brief “patient activation” intervention designed to set expectations and prepare families of adult intensive care unit (ICU) patients to communicate effectively with the clinical team. DESIGN: Phase I study of acceptability and immediate side effects. SETTING AND PARTICIPANTS: 122 healthcare proxies of 111 consecutive patients with a stay of ≥24 hours in the Johns Hopkins Hospital Medical ICU (MICU), in Baltimore, Maryland. INTERVENTION: Reading aloud to proxies from a booklet (Flesch-Kincard reading grade level 3.8) designed with multidisciplinary input including from former MICU proxies. RESULTS: Enrolled proxies had a median age of 51 years old with 83 (68%) female, and 55 (45%) African-American. MICU mortality was 18%, and 37 patients (33%) died in hospital or were discharged to hospice. Among proxies 98% (95% CI: 94% - 100%) agreed or strongly agreed that the intervention was appropriate, 98% (95% CI: 92% - 99%) agreed or strongly agreed that it is important for families to know the information in the booklet, and 54 (44%, 95% CI 35%– 54%) agreed or strongly agreed that parts of the booklet are upsetting. Upset vs. non-upset proxies were not statistically or substantially different in terms of age, sex, education level, race, relation to the patient, or perceived decision-making authority. CONCLUSIONS: This patient activation intervention was acceptable and important to nearly all proxies. Frequently, the intervention was simultaneously rated as both acceptable/important and upsetting. Proxies who rated the intervention as upsetting were not identifiable based on readily available proxy or patient characteristics.