Using EMR data to determine origin of HAI

It's known that certain settings like healthcare, long-term care, and prison facilities increase the risk for certain infections. A prime example is MRSA. The challenge has been to determine where the patient acquired the infection, either from the community at large or from a healthcare facility. The distinction between the two is important for the monitoring, treatment, and prevention of infections.

A study by JS Wilson developed a process that identifies healthcare facilitiese from patients' addresses in the EMR database, automatically categorizes the address as either a healthcare facility or a residential address, and finally decides whether or not the infection is likely to be community or healthcare acquired.

This is an interesting attempt to use EMR data in assessing infection risk and opens the door to further investigating the effects of social networking on infections.

Full research paper here

What's better for your health, "long ties" or "dense cluster?

Lately, there has been a lot of buzz on using social networks to change
health behavior.  Ron blogged earlier about Nicholas Christakis' book
Connected, this week a new research study came of out of MIT that shows that when it
comes to changing health behavior, having fewer friends that one knows
really well is better than having many friends whom one doesn't know well.

Professor Damon Centola from the MIT Sloan School of Management tracked
the number of people who registered for a health forum from two distinct
social networks.  In one social network, participants had "long ties" with
each other, meaning each participant knew many different people but didn't
know them well.  In the other social network, participants formed "dense
clusters," meaning each of them knew fewer people but knew them very well.

The study result showed 54% of the people from the dense clusters network
registered for the health forum and 38% from long ties network did.  This
study suggests that policies may be more effective when aimed at
communities and groups that act as clustered networks.

More about Professor Centola's study