Fig. 9

Perceived dangers of artificial intelligence in setting alarm thresholds by profession

Respondents’ answers to the question, “What dangers do you see when artificial intelligence suggests alarm profiles (e.g., for COPD) on a patient-specific basis and continuously reevaluates them?” Percentages indicate the proportions of answers given per Likert item (e.g., “Agree”) by professional group. Responses marked with an asterisk (✱) are statistically significant regarding the observed median and 95% confidence interval of all responses using bootstrap resampling procedure. Responses marked with a triangle (▲) reveal a statistically significant difference in responses stratified by profession (nurses, doctors) using Mann–Whitney U test (p < 0.05).