Information technologies are transforming nearly every aspect of our lives, and that includes patients’ interactions with health care systems and the delivery of health care itself.
Our investigators are tapping into the potential of health IT to provide needed services to larger numbers of people, at a time and place that is more convenient for them, and at lower cost.
Whether through web-based self-help programs, phone calls using interactive voice recognition, or text messages to study participants, we are demonstrating that health IT innovations can deliver significant, cost-effective benefits at all levels of the health care continuum—from patients to providers to health systems. We are also using sophisticated data-extraction techniques that fully reveal the wealth of information contained in electronic medical records.
At the heart of many of our studies is a tremendous resource: KP HealthConnect, Kaiser Permanente’s electronic medical record. All aspects of our members’ health care—including vital statistics, hospital stays, Emergency Room visits, pharmacy records, mental health care, imaging results, and lab tests—are captured in KP HealthConnect and associated with the member’s unique health record number. Unless they opt out, our health plan members agree upon joining to allow their records to be used for research. The link between our research databases and KP HealthConnect enables our researchers to conduct both retrospective (backward-looking) and prospective (forward-looking) studies.
Our researchers have used KP HealthConnect to assess screening services, monitor quality of care, and identify members who might benefit from new programs. They have also linked it with other technologies, such as automated mailing and calling programs.
Natural Language Processing
When analyzing large numbers of medical records, it can be difficult to observe all the nuances of health and clinical care. Many events of interest to researchers are documented only in the “free text” notes that care providers enter into the medical record. Our researchers have found a way to unlock this data using Natural Language Processing (NLP) technology. This work is led by Brian Hazlehurst, PhD, a cognitive anthropologist who has developed innovative methods for extracting data from written notes in electronic medical records. Dr. Hazlehurst’s team has created a centralized website, CER Hub, where researchers can build software applications that use NLP for processing medical records. The resulting datasets can be compared to facilitate research across many different health care settings.