Machine learning for risk stratification in the emergency department (MARS-ED) study protocol for a randomized controlled pilot trial on the implementation of a prediction model based on machine learning technology predicting 31-day mortality in the emergency department

Table 2 Questionnaires regarding clinical intuition and medical treatment changes

Questionnaire regarding clinical intuition (both intervention group and control group)
- Surprise question: “Would you be surprised if this patient dies within the next 31 days?” (yes/no) [20, 21] - Concern question: “How concerned are you about the health of this patient?” (Likert scale 1–10) [6] - Severity question: “How severely ill do you find this patient?” (Likert scale 1–10) [22] - “Do you think that this patient will be admitted to the hospital for more than 7 days?” (yes/no) - “Do you think that this patient will be admitted to the ICU?” (yes/no)
Questionnaire regarding medical treatment changes (intervention group only)
- “Is the RISK^INDEX higher than, equal to, or lower than you expected?” - “Do you want to alter the medical treatment plan based on the RISK^INDEX?” (yes/no) - “If yes, which part of the treatment plan?“ (reassessment during current ED visit; order additional investigations and consultations; admission or discharge; consultation with an ICU specialist and/or admission to ICU; removing or emplacing treatment restrictions) ^a

ED, emergency department; ICU, intensive care unit
^a The physician will be asked to specify which part of the treatment plan was altered by choosing any of the given options

ISSN: 1757-7241