A vision for patient-centered health information systems.

THE HEALTH INFORMATION TECHNOLOGY MOVEMENT FOcuses much of its energy on the use of electronic medical records by clinicians, but the use of information technology by patients carries equal promise. Outside of health care, the public routinely uses computers and smart phones to access information and perform tasks with a click of a button. Patients seek similar ease in accessing health information, but health care has been slow to develop information tools for patients of comparable functionality. A technology that could respond to this need is the personal health record. A model for shifting personal health records to become patient centered, which stratifies these interactions by arbitrarily defined levels, is depicted in the FIGURE. Initially, personal health records were merely an electronic substitute for the home medical file. The simplest function of the personal health record is to store similar information, often entered by the patient (first level). The user may be asked prompting questions about health behaviors and diagnoses, but the answers come mainly from the user’s memory or home records. Although patients are the ultimate authority on some details, such as eating habits and symptoms, they may not be as precise about diagnoses, medications, and laboratory values. More advanced personal health records address this problem by linking electronically to clinical information in electronic health records or claims data (second level). These systems provide a portal to clinical information, but too many deliver unmodified content to the patient, who may have difficulty understanding the terminology or putting the information in context. For example, a patient may not know whether to be concerned about a creatinine level outside of the normal reference range. A higher function of personal health records is to interpret content—to explain technical information in language that patients easily understand (third level)—or to render clinical advice (fourth level), as when personal health records call attention to overdue screening tests or the need to reduce blood pressure or serum lipid levels. The highest level of functionality is to help patients take action (fifth level). Information is plentiful on the Internet, but some personal health records give patients vetted information, such as a health encyclopedia or hyperlinks to useful resources. Some personal health records personalize information for the individual patient and some can incorporate motivational messages to help patients take action to confront challenges, such as weight loss or smoking. The personal health record offers a platform for applications, such as Framingham or Gail Model calculators or decision aids that help patients weigh difficult tradeoffs. By assisting with self-management, offering logistical support for appointments, and providing follow-up (eg, reminders), personal health records can extend care beyond the clinical encounter. This must be done in coordination with the patient’s physician, a level of functionality that many standalone systems lack. Trust comes when the personal health record has the endorsement of the physician and is designed to interface seamlessly with care delivery, as when output is shared and the personal health record refers patients back to their physician for assistance or helps them prepare for upcoming office visits. No existing products do all of this. Relatively few patients use personal health records, perhaps because most products only have firstor second-level capabilities. Higher functionality exists in some high-end systems, but even these cannot deliver the full spectrum of services, and their developers report slow adoption. The typical personal health record takes an oversimplified approach, such as issuing a blanket reminder for mammograms without considering the many factors that influence whether screening is indicated. The power of smart technology enables personal health records to be far more refined and thereby more appealing to patients. Personal health records that lack this sophistication are potentially harmful at every level of functionality. They can accept erroneous information from patients at the first level or from electronic health records at the second level. Patients who view unmodified content through a portal may be confused or experience undue anxiety without adequate context. At the third and fourth levels, the system’s explanation of findings and clinical advice may be spurious, depart from evidencebased guidelines, or include generic advice—correct for the general population, but wrong for the individual patient. At