The promise of big data has been shimmering on the health care horizon for years. When properly aggregated and accessed, big data has the potential to personalize medical decision-making by identifying which treatments have the best chance of success for an individual patient. Now, athenahealth and Epocrates have come together to bring that idea to reality, in a new app called Epocrates Bugs + Drugs. And here’s how we did it…
An urgent need for infectious disease decision support
The interplay between population health and personalized medicine is poignantly manifest in infectious disease management. One clinician’s antibiotic choices for one patient can result in spread of infection—and resistance—reaching far beyond the exam room. The Centers for Disease Control (CDC) recently reported that more than two million patients suffer from antibiotic-resistant infections each year, resulting in at least 23,000 deaths.
We know the solution is in our hands: hands that deliver self-care instructions instead of antibiotic prescriptions for viral and self-limited conditions; hands washed meticulously after every patient; hands whose fingers scroll and tap to access the latest CDC, IDSA and AAP guidelines.
When it comes to choosing antibiotics, clinicians in hospitals have long benefitted from antibiograms that detail resistance patterns of the bacteria sampled from patients within their walls. For years, I carried a well-worn card or printout of institutional sensitivity results in my white lab coat pocket, updating it every year or so. I would visually scan the tiny print to find a prudent antibiotic choice that would get the job done. Yet, it never occurred to me that there was no similar guide for the ambulatory setting.
Our inspiration for the app
Like most great ideas, the inspiration for our recently announced Epocrates Bugs + Drugs mobile health (mHealth) app came from clinicians who rely on Epocrates in daily practice. They loved using Epocrates’ ID Treatment Selector, but they had to look elsewhere for antibiotic sensitivity data.
The Epocrates Medical Information team gathered to brainstorm a solution and immediately thought about those hospital cards. How could we gather that data? With a great number of infections managed outside the hospital, we knew we had to find a way to access the same type of sensitivity data in the outpatient world, where patients actually live.
We went on a data hunt
Independent labs, hospital systems, various electronic medical record platforms, national surveillance systems—everyone seemed to have a piece of data on antimicrobial sensitivity, but it was all locked in separate silos.
Now, that’s changed. As just one result of Epocrates and athenahealth joining forces earlier this year, we now have access to data from more than 40 million patient records (15 million of which have full clinical data), via the athenahealth cloud-based network. By aggregating and de-identifying individual lab reports received by athenahealth from a variety of labs across the nation, we are able to harness data to create something entirely new: a community-based antimicrobial susceptibility matrix focused on any ZIP code in the country.
From the moment we said, “Let’s do this!” a working prototype was in our hands two weeks later. The app is simple.
How it works:
- Bugs + Drugs geo-locates your Apple® iOS 7 device and delivers a rank-ordered list of bacteria found in urine, blood, or skin specimens predominant in your selected area.
- Depending upon the location you’ve selected, the radius dynamically expands to deliver meaningful results for that area.
- Tap on a bacterium to reveal sensitivity data by drug. Samples with fewer than 30 tests are less reliable, and are presented separately.
- Tap on a drug to open an excerpt from the top-ranked Epocrates drug reference.
- Thousands of bacteria-to-drug data points tied to patient-level ZIP codes are available each day, enabling automatic continual app updates.
Check out Epocrates Bugs + Drugs app in action:
The results are endlessly fascinating
Amoxicillin/clavulanate might be a reasonable empiric choice for a skin infection in most parts of the country, but not if your patient lives in, say, Alexandria, Virginia, where Staph aureus resistance to the drug is worse than the national average. If you have a lab report showing Enterococcus faecalis in urine and you are awaiting sensitivities, nitrofurantoin might be a reasonable choice—but not in Jennings, Louisiana.
Of course, patient-specific factors, clinician judgment, affordability, and other factors are vital to any prescribing decision. Still, the application of geo-specific sensitivity profiles represents another step toward personalized medicine and population health management. The decisions made in the moment of care—whether to prescribe an antibiotic and the choice of drug—can affect an entire population.
So are you ready to use big data to deliver more personalized care? Download the Epocrates Bugs + Drugs app for iOS 7 today. Tell us other ways you think mHealth may be able to improve the future of health care. What other kinds of data applications would you like to see us undertake in the future?