WASHINGTON - The doctor doesn't think your sore throat is bad enough yet to order a strep test - unaware that a dozen people across town were diagnosed with strep throat just last week.

Doctors rarely know what bugs are brewing in the neighborhood until their own waiting rooms start to fill. Harvard University researchers report Monday that getting them real-time information on nearby infections could improve patient care - for strep throat alone, potentially helping tens of thousands avoid either a delayed diagnosis or getting antibiotics they didn't need.

"The risk you have is based on where you live and what the people around you have," explains Dr. Kenneth Mandl of Children's Hospital Boston, affiliated with Harvard. His analysis of 82,000 patient visits found that knowing how much strep throat is circulating can help improve the accuracy of the next patient's diagnosis.

Today, hundreds of hospitals, clinics and health departments automatically report certain symptoms and diagnoses to the government. That practice has a wonky name - biosurveillance - but it's how officials track the spread of flu, detect the latest whooping cough outbreak, and watch for weird symptoms that might signal a brand-new disease or even bioterrorism.

But until there's an outbreak, that information is a one-way street. There's no easy way for doctors to learn what their colleagues nearby are diagnosing. Instead, doctors often call the health department to ask if anyone has heard of a case of this or that disease as they puzzle over a patient's symptoms, says Dr. Alfred DeMaria of the Massachusetts Department of Public Health.

Giving doctors a fast, ongoing snapshot of disease "would be very helpful," says DeMaria, who wasn't involved in Mandl's research but praises the approach. "The key is to make the system entirely automated and real-time."

Work is beginning on technology to do just that, trying to link local biosurveillance to electronic health records, maybe even mobile apps.

First, the question is whether such tracking could make a real difference. So Mandl and his colleague, Dr. Andrew Fine, an emergency medicine physician, examined strep throat, an infection frequently misdiagnosed in adults.

Because strep throat is more common in young children, those with red, sore throats are given either a while-you-wait rapid test or, because that test sometimes misses the bug, a throat culture that can take a day or two for results.

For anyone 15 or older, guidelines say doctors shouldn't order a test or prescribe antibiotics unless sore-throat sufferers rise to a certain level of suspicion because of other symptoms: fever, enlarged lymph nodes, tonsils with swelling or pus, and a lack of coughing. People with none or just one of those symptoms probably have a virus and are supposed to be sent home. A patient with a lot of those symptoms often is given antibiotics automatically, and those in between get tested.

Mandl and Fine turned to records from CVS MinuteClinics in six states where all sore-throat patients are tested and symptoms are recorded. What the government-funded study found: Knowing how prevalent strep is in a particular area is a strong enough predictor to count as an extra symptom in the test-or-treat decision.

In other words, if little strep throat is circulating, the chances that someone with several strep symptoms really has it drops enough that it's worth testing them before prescribing antibiotics. Considering there are 10.5 million annual health-care visits for suspected strep throat, that change could prevent unnecessary antibiotics for more than 166,000 patients, the researchers reported.

On the flip side, someone with just a sore throat and fever usually wouldn't get tested but if the strep germ is prevalent in their community, testing could spot 62,000 previously missed patients nationally, the researchers calculated. Their study appears Monday in the journal Annals of Internal Medicine.

Strep isn't the only example. In smaller studies, Mandl and Fine have found that knowing how much whooping cough and bacterial meningitis are spreading locally can improve diagnosis of those diseases, too.

The challenge is how to disseminate such information fast enough for doctors to use. MinuteClinic, for example, says it doesn't record test results in real time yet but may be able to share that kind of data in the future.

Broadening the concept, federal health officials are working to create an easy-to-use Web tool that would let doctors search for local surveillance information. They also are testing how to automatically send alerts about disease outbreaks to the electronic health records of patients with similar symptoms.

Lauran Neergaard covers health and medical issues for The Associated Press.