Dictionaries for Understanding

Dictionaries for UnderstandingWhen we talk about using the patient’s own language to communicate about health goals, we necessarily don’t mean English vs. Mandarin (although of course that type of language match is important too!). What we really mean is understanding what the patient values and how the patient expresses his understanding of his health, so we can adopt terminology that resonates with him. Understanding the patient’s language also allows us to better detect and comprehend statements about the experience of health.

Dr. John Brownstein, Harvard Medical School, Data and Design for Disease Detection, spoke at the Hx Refactored conference about work he’s done mining social media to understand disease outbreaks. This type of investigation depends on understanding the many different ways people might refer to their health. Consider the flu: Some people may understand that they have the flu and use clinical language to communicate about it, but others might focus on symptoms like coughs and body aches, while others might talk more generically about feeling lousy. Can you then also separate all of these folks from the people who don’t have the flu but speculate they might based on other symptoms? The challenge for researchers here is knowing what type of language people use to talk about health and being sensitive enough to it that they can accurately gauge meaning.

You can imagine that this type of research might lead to literal glossaries or dictionaries to help investigators better understand the world of terminology in use by patients. It reminds me of something I recently learned about; George Matsell, in New York City in the 1850s, wrote a book called The Secret Language of Crime: Vocabulum or the Rogue S LexiconMatsell compiled the slang used by street criminals, known as “flash,” into a dictionary for use by the police.

No doubt, a dictionary is a clumsy translation tool for dealing with patients, but it might be a helpful first step to codify procedures for researchers like Brownstein who are sifting through enormous amounts of qualitative patient-generated data. This type of tool may also have an important role to play in the development of patient facing technology that does naturalistic language processing, like the work coming out of Henry Lieberman’s group at MIT Media Lab focusing on artificial intelligence and “common sense computing.”