Hospitals collect discharge data from all of their patients -- it's mostly used for billing purposes. A Drexel University public health researcher says this readily available data should be mined to help prevent medical accidents.

Jennifer Taylor analyzed discharge data from all Pennsylvania hospitals for a year -- 1.9 million patients. She found that almost 10 percent experienced a so-called "patient safety event," meaning they got an infection, were given the wrong medication, experienced a surgical error or other problems.

She says hospitals underutilize discharge information in terms of preventing medical errors and accidents.

"It's very powerful data," explained Taylor.  "We use it a lot in public health, the state health department uses it a lot, what we are saying in this paper is that hospitals can use it too."

Taylor says discharge data sheds light on which types of mistakes and accidents are most common, and how they could be prevented.

"Like giving a person the wrong drug, or having a person fall while they are in the hospital, those are things we know can be prevented," she said. "Then we should be looking at that data, and hospitals should be asking themselves, and working with public health researchers in how do we prevent this from happening."

Taylor says the data also shows which patients are the most vulnerable. For example, she found that patients over 70 were more likely to experience adverse medication events.