How reams of geographic health data can remake medicine

BALTIMORE  -  Imagine a clinical database containing medical histories, diseases and demographic data taken from millions of patients across the U.S. Imagine data analytics that could help eliminate healthcare disparities and socioeconomic inequalities, could help pinpoint certain diseases in particular populations, reduce morbidity numbers, and rather than improve healthcare for a single patient, could potentially improve the healthcare of millions across a geographic area. 

Enter the up-and-coming healthcare star: population health.  

Numerous organizations and education institutions are trading the traditional "one patient, one doctor" approach for the "entire population approach." 

The most recent newcomer to this burgeoning field is the Johns Hopkins Center for Population Health IT (CPHIT), which was started in July of this year. 

Officials at the center cite population health as an often more effective way of treating patients and geographic groups.  

"Those of us in public health believe that the true way to add value in medical care is to think about the wellbeing of the entire population," says Jonathan Weiner, professor in the Department of Health Policy and Management at the Johns Hopkins Bloomberg School of Public Health and director of CPHIT. And part of adding that value, he added, requires utilizing e-health. 

 "We are currently witnessing the most expansive digitalization of healthcare in history," says Weiner. "Over the coming decade, electronic health records, personal health records and other e-health applications will completely transform healthcare in the U.S. and around the world. Today the focus is clinicians using these systems to treat the individual patient. Our Center wants to extend this focus to enable public health agencies and accountable provider or payer organizations to also harness these health IT systems to create solutions for the many population health issues facing our nation."   

Although in its early beginnings, CPHIT officials anticipate the center will lead the way in population health IT come near future. 

Currently, CPHIT does not have independent projects established but is working in collaboration with colleagues across Johns Hopkins. 

Weiner says the center is working in conjunction with scientists at the Center for Language and Speech Processing at Johns Hopkins, on a project focused on outreach for high-risk moms for the university's Medicaid health plan. Because many of the nurses go through the terms manually, or many doctors currently use free text, the Natural Language Processing (NLP) scientists are developing algorithms to do this more efficiently.   

Moreover, CPHIT is also working on a project for the Centers for Medicare and Medicaid Services (CMS), helping the agency understand how electronic health records can be used for advanced quality metrics as part of meaningful use, Weiner added. "It's easier said than done to take quality measure out of an electronic medical record because no two records are alike, and no two doctors or nurses use it in the same way." 

However, Weiner says today the best data does not come from a patient's electronic medical record, but rather from insurance claims data, which he calls the "backbone" of clinical computerized data. "[Claims data] may not be perfect, but they're the source for diagnostics and pharmaceutical information in the U.S. right now; it's not the EMR," he adds. In the future, he says, registries will be the next big 

A project at Boston Children's Hospital is also embracing population health IT, but on a smaller scale than what Weiner and his colleagues are doing. Take the Informatics for Integrating Biology and the Bedside (i2b2) for instance. Established by the National Institute of Health and a National Center for Biomedical Computing based at Partners HealthCare System, is doing what population health does on a slightly smaller scale. 

The project uses existing clinical data to find health solutions for groups of patients. Marvin Harper, CMIO at Boston Children's Hospital, calls this population health management a new category of medicine.

 "So you can define your population by a drug they take, an allergy they have, by a disorder they have and be able to suddenly both identify them and then track certain characteristics about them, which will tell you how well or how poorly you're doing in managing that population," Harper says. 

Daniel Nigrin, CIO at Boston Children's Hospital adds that this i2b2 population health tool can search hundreds of thousands of patients, and takes only a matter of seconds based on the criteria you search for. 

"Say we find that there's all of sudden a medication that's found to have an adverse effect, because we've got medications now in our system," Nigrin says, "We can ask a tool like this to find us all of the patients who are on medication X that we know of and quickly get a list and contact them." 

Population health IT experts say the future will only bring more promises for clinical care. "One day all digital data can and will be integrated into e-health," Weiner says. The implications of this for patient health, he adds, are overwhelming.