OBJECTIVES: The main hypothesis in this study is that stakeholders have different preferences concerning IT innovations in hospitals, and these preferences are caused by perceived cost/benefit ratios. This will translate in disagreement between stakeholders on which innovations to implement first, possibly explaining the slow diffusion of innovations in health care. METHODS: Analytic Hierarchy Process (AHP) was used to quantify stakeholders positions in their priority of nine IT innovations. These innovations were selected after a systematic literature review and expert interviews. In the AHP, decision criteria related to costs and benefits of the innovations were defined: improvement in efficiency, health gains, satisfaction with care process, and required investments. Stakeholders judged the importance of the decision criteria and prioritized the selected IT innovations according to their expectations of how well the innovations would perform on these decision criteria. RESULTS: Sixty-two respondents, including patients, nurses, physicians, managers, health care insurers and policy makers showed significant differences in their expectations about their respective costs and benefits of the innovations, resulting in diverging preferences for the health care innovations. For instance, self tests are one of the most preferred innovations by health care insurers and managers, due to its expected positive impacts on efficiency and health gains. However, physicians, nurses and patients strongly doubt the health gains of self tests, and accordingly rank self tests as the least preferred innovation.  CONCLUSIONS: We found clear differences in expectations of different stakeholder groups on IT innovations. The differences can be understood from the perspective of costs and benefits per stakeholder for each innovation. This study gives a first quantitative insight in stakeholder differences and presents a novel way to study stakeholder differences. The results may be used by decision makers to include alignment of stakeholder positions in implementation processes.