BACKGROUND: The number of physician-rating websites (PRWs) is rising rapidly, but usage is still poor. So far, there has been little discussion about what kind of variables influence usage of PRWs. OBJECTIVE: We focused on sociodemographic variables, psychographic variables, and health status of PRW users and nonusers. METHODS: An online survey of 1006 randomly selected German patients was conducted in September 2012. We analyzed the patients' knowledge and use of online PRWs. We also analyzed the impact of sociodemographic variables (gender, age, and education), psychographic variables (eg, feelings toward the Internet, digital literacy), and health status on use or nonuse as well as the judgment of and behavior intentions toward PRWs. The survey instrument was based on existing literature and was guided by several research questions. RESULTS: A total of 29.3% (289/986) of the sample knew of a PRW and 26.1% (257/986) had already used a PRW. Younger people were more prone than older ones to use PRWs (t967=2.27, P=.02). Women used them more than men (χ(2) 1=9.4, P=.002), the more highly educated more than less educated people (χ(2) 4=19.7, P=.001), and people with chronic diseases more than people without (χ(2) 1=5.6, P=.02). No differences were found between users and nonusers in their daily private Internet use and in their use of the Internet for health-related information. Users had more positive feelings about the Internet and other Web-based applications in general (t489=3.07, P=.002) than nonusers, and they had higher digital literacy (t520=4.20, P<.001). Users ascribed higher usefulness to PRWs than nonusers (t612=11.61, P<.001) and users trusted information on PRWs to a greater degree than nonusers (t559=11.48, P<.001). Users were also more likely to rate a physician on a PRW in the future (t367=7.63, P<.001) and to use a PRW in the future (t619=15.01, P<.001). The results of 2 binary logistic regression analyses demonstrated that sociodemographic variables (gender, age, education) and health status alone did not predict whether persons were prone to use PRWs or not. Adding psychographic variables and information-seeking behavior variables to the binary logistic regression analyses led to a satisfying fit of the model and revealed that higher education, poorer health status, higher digital literacy (at the 10% level of significance), lower importance of family and pharmacist for health-related information, higher trust in information on PRWs, and higher appraisal of usefulness of PRWs served as significant predictors for usage of PRWs. CONCLUSIONS: Sociodemographic variables alone do not sufficiently predict use or nonuse of PRWs; specific psychographic variables and health status need to be taken into account. The results can help designers of PRWs to better tailor their product to specific target groups, which may increase use of PRWs in the future.