To trust or not to trust, that is the question: AI-supported screening in general practice – a cross-sectional study on trust among urban and rural citizens (Preprint)

Abstract

<sec> <title>BACKGROUND</title> The early detection of diseases is one of the tasks of general practice. AI (artificial intelligence)-based technologies could be useful to identify diseases at an early stage in general practices. As a good 90% of the population regularly consult a GP (general practitioner) during one year, this could increase the percentage of citizens who take part in meaningful screening measures. </sec> <sec> <title>OBJECTIVE</title> Considering these factors, the aim of the study was to evaluate the level of trust of citizens in rural and urban areas in AI-supported early detection measures in general practice. </sec> <sec> <title>METHODS</title> This cross-sectional study was conducted in Schleswig-Holstein, Germany from Nov 2023 to Dec 2023 on the topic of early detection measures with AI in general practice care, among other things. For this purpose, 5,000 adult residents of rural areas (Ostholstein, Pinneberg, Nordfriesland) and urban areas (Kiel City) were invited to take part in the survey. Data analysis was carried out using descriptive statistics, sub-group analysis, linear and stepwise regressions to identify the factors that influence trust in AI-based diagnoses. </sec> <sec> <title>RESULTS</title> The majority of respondents 55% (n=787) consider the introduction of an AI-based screening measure to be a sign of modern medicine. Moreover, 27% (n=388) of respondents fear that the introduction of such services could lead to a deterioration in the doctor-patient relationship. The role of AI in future care was rated as (very) important by 35% (n= 634). The stepwise regression analysis showed that a positive attitude towards AI in medicine being the strongest predictor (ß= 0.420) concerning trust in AI based diagnoses. In contrast, trust in physician diagnoses was associated with lower age (ß= -0.111) and shorter waiting times for test results (ß= 0.077). </sec> <sec> <title>CONCLUSIONS</title> Trust in a GP based diagnose was around six times greater than towards AI applications. Despite concerns about their impact on the doctor-patient relationship, a good third of participants believe that the role of AI in healthcare will grow. Interestingly, trust in doctors appears to be less pronounced among citizens than among patients. </sec>

Ähnliche Arbeiten