PS 4-5
INTENSIVE CARE UNIT CLINICIANS' PERCEPTIONS OF PROGNOSTIC ACCURACY FEEDBACK
Purpose: Intensive care unit (ICU) survivors and ICU surrogate decision makers want their clinicians to provide prognostic guidance to help inform their preference-sensitive decisions. However, ICU clinicians struggle to communicate this information about mortality and functional outcomes, partially because of their discomfort with prognostic uncertainty. Providing clinicians with feedback about their prognostic accuracy and strategies to improve it may increase the likelihood that clinicians will discuss prognostic information. To explore this hypothesis, we evaluated clinicians' perceptions of the utility of such feedback for critically ill patients.
Methods: We conducted semi-structured interviews with ICU physicians and nurses who provided prognostic estimates of 6-month mortality and functional outcomes for ³5 patients they cared for in the ICU as part of prospective study of prognostic accuracy. We provided each clinician with an individualized prognostic accuracy report and asked them to respond with (1) their receptiveness to such feedback and (2) the impact such feedback may have on their patient care or communication practices. We analyzed responses using constant comparative methods to identify and categorize emerging themes.
Results: 10 eligible clinicians completed the interview and thematic saturation was reached. We identified three major themes (Table): (1) clinicians expressed greater confidence in prognosticating about short-term mortality than long-term functional outcomes (i.e. toileting independently); (2) clinicians felt unprepared to predict 6-month functional outcomes, but recognized the value of discussing future functional status with critically ill patients and family members; (3) clinicians believed that prognostication feedback would enable them to calibrate their prognostications, thus improving their accuracy and increasing their willingness to provide such information to ICU patients and surrogate decision makers as a component of shared decision making (SDM).
Conclusions: Our findings suggest that clinicians have strong desires to prognosticate about long-term functional outcomes while engaging in SDM but lack confidence in their abilities to provide accurate information. Providing clinicians with feedback regarding their prognostic accuracy, particularly in a manner that emphasizes targeted areas for improvement, may increase their accuracy and willingness to discuss functional outcomes with ICU patients and surrogates. Future research should focus on exploring effective and scalable ways to provide such looped feedback to clinicians in order to improve SDM.