JMIR Publications

ABSTRACT

Background:

The growing integration of Personalized Risk Prediction (PRP) and Artificial Intelligence (AI) substantially re-shapes diagnostic and therapeutic decision-making in health care. At the same time, its responsible adoption depends not only on technical performance, but also on patients’ perspectives and acceptance.

Objective:

This study systematically examined patients’ perspectives across several European countries and explored how patients’ technology-related attitudes relate to their evaluations of personalized and AI-supported ap-proaches in cardiac care. As part of the PROFID (Prevention of Sudden Cardiac Death After Myocardial Infarc-tion by Defibrillator Implantation) project, its focus is on the ethical use of PRP and AI in the clinical context of decision-making regarding sudden cardiac death (SCD) prevention and implantable cardioverter-defibrillator (ICD) implantation.

Methods:

The study used a cross-sectional survey design with a standardized questionnaire including multimedia con-tent. The target population comprised adults aged 18 years or older living in six European countries who met at least one of the following (self-reported) clinical criteria: heart failure, myocardial infarction (MI), cardiac arrest, or current ICD implantation. An exploratory factor analysis (EFA) was used to identify and evaluate internally consistent scales, and subsequent regression analyses examined associations between these scales and technological openness, sociodemographic characteristics, and patients’ views on PRP and AI in cardiac care.

Results:

The sample consisted of 470 participants from Germany (n=210), the Netherlands (n=86), the United Kingdom (n=145), and three other European countries (n=29; Austria, Belgium, and Spain). Overall, 51.9% (244/470) of respondents were male and 48.1% (226/470) were female. The mean age of the sample was 61.12 (SD 12.62) years. The EFA showed six clearly interpretable factors: (1) Perceived benefits and support of PRP models in medical decision-making (MDM); (2) Perceived benefits and support of AI in MDM; (3) Transparency expectations in algorithmic decision-making; (4) Support for delegating decisions to algorithms; (5) Self-reported AI literacy and (6) Preference for shared decision-making (SDM). The regression analysis showed the relations of technologi-cal readiness, self-reported AI literacy, support for delegation of decisions to algorithms, transparency expecta-tions in algorithmic decision-making, preferences for SDM, and educational attainment to predict patients’ perceived benefits and support of PRP or AI in MDM.

Conclusions:

The findings support existing assumptions while also highlighting additional aspects that should be considered if high-level technologies are used in decision-making processes related to ICD implantation. PRP and AI were generally perceived as useful tools to support decision-making regarding ICD indication, provided that trans-parency is ensured and patients remain actively involved in the decision-making process. Mandatory use and full delegation to decision-making directly by Al were broadly rejected. Generally, men showed more positive perceptions of the use of AI in MDM than women. The attributed acceptance of delegation to PRP models was significantly higher than AI.