Personal health records (PHRs) are electronic medical records whose contents are controlled and managed by the consumers whose data they carry. They have attracted much press attention in recent years, and both industry and governments in many countries see them as potential solutions to many IT problems in health care.? ? The National Library of Medicine (NLM) has embarked on the development and deployment of a Personal health record (PHR) in order to study and improve their utility, reduce the barriers to their use, identify best practices, and provide a platform and test bed for advanced applications. The development is based on a set of existing messages and vocabulary standards that are supported by the Health and Human Services and NLM and on both existing (Ruby on Rails, Scriptaculous, and Dojo) and NLM-developed open source software building blocks, some of which will have broad medical informatics applications. NLM's PHR is a pure web application. It is based on a forms generator (developed by our group) that produces web input forms on the fly. These forms include built-in skip logic (e.g. if the person is male, it does not ask about pregnancy history), edit checks and auto complete input. It uses AJAX techniques, so the response time is very fast. ? ? The NLM PHR provides tools for managing clinical information, from many different family members, so a mother can maintain immunization records for each of her children and/or keep track of her ailing fathers medications. It provides places for recording medications, medical problems, surgeries, immunizations (vaccines), important measurements such as blood pressure and laboratory results (e.g. serum glucose). It is designed to encode the names of the major facts, e.g. drugs, problems, surgeries, immunizations, it carries. The encoding translates names into federally supported coding systems, e.g. Rx.terms (a Subset of Rx.Norm developed for this project and adopted by The Centers for Medicare and Medicaid Services (CMS) for their post-acute care project) for drugs, Logical Observation Identifiers Names and Codes (LOINC) for measurements, the Centers for Disease Controls (CDC) vaccine codes for immunizations, Systemized Nomenclature of Medicine (SNOMED) for diagnoses. It adopts Health Level 7 (HL7) for much of its data structure and all of its data types. Both of these capabilities facilitate the importation of clinical data from widely available HL7 messages.? ? The orientation toward automatic coding enables NLMs PHR to provide two special capabilities: 1) decision support rules and 2) one-click access to information about concepts, e.g. symptoms, medications, immunizations, recorded in the record.? ? Decision support in the NLM PHR is based on predefined rules that tie patient conditions to care recommendations. The NLM PHR compares the patients data against rules for preventive care and reminds the individual about interventions such as mammograms or colonoscopy which are due. ? ? A button appears next to every coded concept on the screen. These buttons provide one-click access to consumer-tailored information about the concept. For medical problems and drugs the information source is NLMs MedlinePlus, for immunizations the CDCs vaccine information, for preventive care reminders it is US Preventive Task Force web site, etc. When you click on a button next to a concept, the computer immediately pops up 2-4 pages of information.? ? We envision that the NLMs PHR is being used by individuals and family caregivers. ? ? This is a very young project to produce something that has been a longstanding NLM interest and is part of NLMs strategic plan. It will use and tune the message and vocabulary standards that NLM has supported and also provide another consumer entry point to NLMs rich trove of patient-oriented data. Early research projects will be focused on user's needs, usability, and usage patterns to guide the next round of development and research.
