Federal incentives to create a nationwide system of electronic health records (EHRs), together with advances in information technology, are rapidly leading to the ability to identify cohorts of patients with precise attributes. This process, known as "EHR phenotyping", applies high-throughput algorithms to electronic data to classify patients based on exact constellations of information (e.g., demographics, diagnoses, procedures, laboratory values, vital signs, medications, lifestyle and environmental factors). EHR phenotyping is expected to result in studies with greater power and lower costs, and is a key component of the vision for learning healthcare systems that support an array of clinical, observational, outcomes, and comparative effectiveness research. The ultimate success of this enterprise depends on building and maintaining public trust, and patient input is vital. Little is known about patients'willingness to share their data for research purposes, their preferred level of control over such use, or their perspectives on the need for and acceptability of different approaches to informed consent. In addition, EHR data are far from perfect, reflecting the noise and complexity inherent in the healthcare system and thus subject to incompleteness, inaccuracies, and bias. Researchers using EHRs will almost certainly uncover discrepancies (e.g., between diagnosis codes and lab values), and find themselves in the position of needing to contact patients-either to inform them of a serious potential health concern, or to otherwise resolve the discrepancy. This is a novel challenge that researchers will increasingly confront. The objective of the proposed research is to help fill these gaps by gathering empirical data from patients in four highly diverse counties in the southeastern US, capitalizing on two existing studies taking place in these counties to obtain rich, policy-relevant data both on patients'opinions and their actual behavior. To attain this objective, we will: (1) Conduct semi-structured interviews to assess patients'willingness to share their clinical data for research use, including acceptable approaches to informed consent;(2) Investigate patients'reactions to researcher contact based on the results of EHR phenotyping, through focus group research as well as analyses of audio recordings of actual telephone calls made by researchers to participants to resolve discrepancies between participants'self-reported health information and their EHR data;and (3) Convene a series of deliberative democracy events to systematically develop recommendations concerning consent and contact with patients to inform the ethical use of clinical data for research.
The widespread adoption and use of electronic health records will make an unprecedented amount of information available for health-related research. Although clinical records have long been used for research, the scope of the information now available, and the scale on which it may be linked and used, raises urgent issues concerning privacy, confidentiality, informed consent, and the discovery of discrepancies in the data necessitating patient contact. The long-term goal of this research is to support the development of ethical approaches to research use of clinical data that enhance public trust and facilitate scientific progress.