Background: Involving patients as partners in research is a defining characteristic of patient-centered outcomes research (PCOR). While patients’ experiential knowledge of a health condition or treatment may yield research priorities not reflected by researchers and policy makers, the methods for identifying and effectively collaborating with patients are still evolving. Patient registries and crowdsourcing may offer ease of access and convenience to both researchers and patients. Surveys and focus groups, including online modalities, have been described for prioritizing research topics. However, little is known about how these different methods compare in producing consistent priorities and similar perceptions of engagement quality among participants. Objective: The aims of this study are (1) to compare how different engagement methods used to elicit patient priorities for research perform as measured by rankings for priorities generated and participant satisfaction; and (2) to determine characteristics of individuals choosing to participate in research prioritization activities. Methods: Participants in the Back pain Outcomes using Longitudinal Data (BOLD) patient registry, established to evaluate the natural history of back pain among individuals 65 years and older, and participants on the Amazon Mechanical Turk (MTurk) crowdsourcing platform, to provide input on priorities for research via a questionnaire, are invited. For BOLD participants, we subsequently randomize interested respondents to 1 of 3 interactive prioritization activities to further develop priorities: a Delphi panel, an online crowd voting activity, or an in-person facilitated prioritization activity using nominal group technique (NGT). Participants involved in each activity complete a survey to evaluate the quality of the experience and a subset of these participants discuss their experience further in an interview. Descriptive statistics are used to characterize the rankings produced by each method and compare the top 5 rated topics resulting from each prioritization activity. We use rank-ordered logistic regression models to identify associations of the ranked priority topics with baseline patient characteristics. We analyze responses to the evaluation using a mixed-methods approach wherein we tabulate responses to Likert-scale questions and use content analysis to enumerate themes emerging from interviews for the 3 activities. Results: In Phase I, we invite approximately 3000 BOLD participants and 500 Amazon MTurk workers to complete a research topic prioritization survey. Based on these results, we include additional topics into a subsequent prioritization survey. In Phase II, we invite BOLD participants to join 1 of 3 activities: 90 participants for Delphi panel, 100 participants for crowd voting, and 60 participants for focus groups. Of the Phase II participants, 30 will be interviewed to evaluate the activities. Conclusions: This study informs decisions about how to conduct outreach to patient registry participants for providing input on research priorities, how individuals 65 years and older wish to participate in engagement activities, and how different research prioritization methods compare in terms of rankings generated and participant satisfaction. [JMIR Res Protoc 2017;6(9):e168]