Testing a Personally Controlled Health Record to Ameliorate Influenza at MIT

The goal of our work is to test an information infrastructure that creates a three-way link among individuals, health care providers and public health authorities, testing the efficacy of this system for improving the prevention and control of influenza. Our model comprises a full feedback system in which an electronic personally controlled health record (PCHR) provides a primary data source for population-based surveillance and a vehicle for delivery of health promotion messages, in this case about influenza immunization and control. This system will feed into and draw on data from a statewide real-time influenza surveillance system, and be integrated with hospital/provider data systems, forming the three-way link. The influenza pandemic is a "great global infectious disease threat" {Osterholm, 2005} for which we are ill prepared. With limited vaccine supply for the past two years and probably next year as well, it is critical to: understand who should be vaccinated to ameliorate population-wide transmission; have effective channels for delivering health directives to high risk groups; monitor efficacy. We will deploy and test our three-way link model and PCHR record system in a multisite study and evaluate using a randomized controlled design the effects on immunization rates of tailored messages about influenza immunization delivered to patients through the electronic health record. Our focus is children ages 6 months to ten years of age and their household members (two of four settings), all age groups including children and working adults (two settings), and university students/staff (one setting). The specific aims follow.

Aim 1.To leverage an electronic personally controlled health record as a public health tool, establishing three way communication systems among individuals, health systems, and public health agencies. We will establish a formal mechanism for patients to integrate and share healthcare-based and personally entered information. Individuals can share information from their medical record, and in response to surveys delivered through the system. There will be two way communications between the record and the State Immunization Registry. Particular attention will be paid to information security, compliance with the Health Insurance Portability and Accountability Act (HIPAA) privacy rule, and personal control. Primary measurements include the structural integrity of the information system, use and usability of the system, and user acceptance. The main hypothesis to be tested is that people are generally willing to share their personally controlled data to promote personal and public health.

Aim 2. To adapt a traditional syndromic surveillance system, initially designed as an early detector of outbreaks, to nimbly identify patients to be targeted for behavior change through personally controlled electronic records. We will develop health messages as cues to action, directed at improving individual compliance with influenza prevention, calibrating our detection models to identify high risk groups, based on demographic and clinical data contained in personal health records. Messages will be targeted at appropriate risk groups and contain information tailored to the individual's characteristics. The primary focus will be to encourage influenza immunization when indicated and conversely, to encourage delay in immunization under conditions of low risk and scarcity. We will change the content, tempo, and targets of the messages based on real time influenza surveillance. The main hypothesis to be tested is that a real-time surveillance system can be used to calibrate health messaging. A second hypothesis is that sentinel individuals can provide data to surveillance systems that adds value to what is already collected.

Aim 3. To measure the impact on immunization rates of tailored health messages, linked to surveillance, and delivered through a personal health record. We will evaluate the impact of directed communications delivered to randomly selected study participants through a personal health record on influenza vaccination rates comparing them to rates among control participants. Secondary outcomes will include message impacts on rates of self-reported illness, knowledge, attitudes and beliefs about influenza and immunization efficacy, use of recommended health protective behaviors. We will investigate the impacts on technology use and immunization rates of a range of individual social demographic characteristics, household characteristics, and computing-related factors that may influence the acceptability of the proposed technology and/or target health behaviors. The main hypothesis to be tested is that individuals exposed to tailored health messages delivered through a PCHR will have higher influenza immunization rates than their unexposed peers. Secondary hypotheses are that compliance will be moderated by computing experience, acceptability/accessibility, individual demographics (e.g., age, sex, race/ethnicity, social class) social and scheduling supports (e.g., work/time schedule flexibility) , family health factors, and logistical factors (transportation, ease of access to immunization sites, convenience, clear scheduling/notification and minimal waiting time for services).