Reduction of patient trust in physicians when supported by artificial intelligence: results of a vignette study among patients

Abstract

Abstract Background Clinical Decision Support Systems (CDSSs) based on routine care data using artificial intelligence (AI) are increasingly deployed. CDSSs, however, may affect patient-physician relationships (Figure 1) and potentially result in a decrease in treatment adherence. Communication about these systems has mostly focused on technical aspects and less on their acceptance and trust. We aimed to investigate whether patient trust in physicians is affected when supported by a CDSS. Methods We conducted a vignette study among the patient panel (N=860) of a University Medical Center. Patients were presented either a high-risk or a low-risk vignette. For both vignettes, a physician made a treatment decision with (intervention group) or without (control group) the use of a CDSS. Using a questionnaire with a 1-7 point Likert scale (1 = strongly disagree, 7 = strongly agree), we collected data on patient trust in physicians in three dimensions: competence, integrity and benevolence, and trust in technology in general. We assessed differences between the control and intervention groups using Mann-Whitney U tests and potential effect modification by sex, age and education level using multivariate analyses. Results In total, 398 individuals participated. Trust in general was high (5.8 in the high-risk case and 6.0 in the low-risk case). Figure 2 illustrates the distribution of the dimensions of trust, by case and group. In the high-risk case, perceived competence and integrity was lower in the intervention group compared to the control group (5.8 vs. 5.6 and 6.3 vs. 6.0, respectively). In the intervention group, perception of the physician’s competence and integrity was significantly lower in women than in men. No differences between the groups were found in the trust dimensions within the low-risk case. Participants with higher trust in technology in general also showed a higher confidence in physicians using a CDSS in terms of benevolence and integrity (P=.009 and P=.044, respectively). Conclusion We found that, in general, trust in physicians was high. However, our findings indicate a potentially negative effect of AI applications on the patient-physician relationship, especially in women and high-risk situations. Trust in technology in general might increase the likelihood of embracing the use of CDSSs by the treating professional.