LuminaCare using big data to pick the right antibiotic for your infection

Here’s a dirty little secret: When you go to the hospital with a bacterial infection and are put on antibiotics, there’s a good chance the doctor who writes the prescription doesn’t know what she’s doing.It’s not that the physician isn’t knowledgeable; it’s just really hard to know right away which bacterium is the culprit, and which drug will kill it. So doctors make a guess or, more often, prescribe a broad spectrum antibiotic in hopes of netting whichever strain has been filling your big toe with pus ever since you got that $6 pedicure at Felicity’s Discount Nail Salon.Boston startup LuminaCare Solutions is working to take the guesswork out of antibiotic prescriptions with predictive software that can tell doctors which medication is most likely to work for each patient.

“There’s a lot of shoot in the dark. No one wants to say that, but yes,” said chief executive David Howe. “What we’re doing is trying to get the right drug on board at the very beginning.”

The effort is promising enough that the company was accepted Monday to the Startup Health Academy, a three-year health care accelerator program that provides coaching and networking in exchange for a small equity stake. Enrolled companies gather for quarterly summits but otherwise remain in their original locations.

LuminaCare is taking advantage of advances in diagnostics that are shortening the time required to identify bacterial strains from two or three days to just a few hours. It’s starting with software that tracks how patients respond to medications, so that doctors can easily see what’s working and what isn’t. The company is hammering out pilot agreements with several hospitals.

So when a 23-year-old white male who weighs 215 pounds and recently returned from a trip to Australia shows up with an apparent infection, his physician will be able to look up the most common diagnosis for patients who fit the same profile, and see which drugs worked.

The predictions won’t always be right, but faster testing will make it possible to see that an antibiotic is ineffective within hours. A doctor could be wrong on the first try, be forced to change treatment courses, and still find the right drug before an old-school culture sample would have identified the bacterial strain.

In cases where the prediction is correct, there are two major benefits: One, obviously, is that the patient receives proper care right away.

The other is that the rest of the hospital’s antibiotics stay in the bottle, helping preserve their power for other patients. The more often an antibiotic is used, the more likely it is that bacteria will develop resistance to it.

Overprescribing antibiotics through trial and error has contributed to the evolution of drug-resistant bacteria. (If you’re up for a good scare, watch the Frontline documentary about superbugs. It’ll frighten the heck out of you.)

LuminaCare views its software not only as a short-term treatment tool but also as a way to extend the lifespan of antibiotics.

Cal Borchers is a business reporter for the Boston Globe.