BACKGROUND: A large literature informs guidance for GPs about communicating quantitative risk information so as to facilitate shared decision making. However, relatively little has been written about how patients utilise such information in practice. AIM: To understand the role of quantitative risk information in patients' accounts of decisions about taking statins. DESIGN AND SETTING: This was a qualitative study, with participants recruited and interviewed in community settings. METHOD: Semi-structured interviews were conducted with 34 participants aged >50 years, all of whom had been offered statins. Data were analysed thematically, using elements of the constant comparative method. RESULTS: Interviewees drew frequently on numerical test results to explain their decisions about preventive medication. In contrast, they seldom mentioned quantitative risk information, and never offered it as a rationale for action. Test results were spoken of as objects of concern despite an often-explicit absence of understanding, so lack of understanding seems unlikely to explain the non-use of risk estimates. Preventive medication was seen as 'necessary' either to treat test results, or because of personalised, unequivocal advice from a doctor. CONCLUSION: This study's findings call into question the assumption that people will heed and use numerical risk information once they understand it; these data highlight the need to consider the ways in which different kinds of knowledge are used in practice in everyday contexts. There was little evidence from this study that understanding probabilistic risk information was a necessary or valued condition for making decisions about statin use.