Monday, May 27, 2013

Data Exhaust

In 2010 more than a billion visits were made to clinicians in the United States by patients in the ambulatory setting. Ambulatory means an office based, outpatient clinic. This is compared to, and specifically not an emergency room, or surgery, or a clinic inside of a hospital, but a free standing building. Most of those patients were seen by general practitioners.

Of those 1 billion encounters approx. 50 million left their visit having been diagnosed with a disease of the skin or subcutaneous tissue (a word meaning just underneath the skin). These numbers of skin disease diagnosis come from visits to pediatricians, family medicine physicians, as well as dermatology specialists.This data is from the publicly available National Ambulatory Medical Care Survey, conducted yearly by the United States Centers for Disease control.

The table below displays results for skin diagnosis for the years 2008, 2009, and 2010.

Primary diagnosis at office visits, classified by major disease category. United states. 2008-2010

Diagnosed as: Diseases of the skin and subcutaneous tissue


Considering the previous post indicating the size of currently active skin image databases, these numbers suggest they are performing at a margin below their capacity. The implicit suggestion here is that at every office visit that a healthcare provider diagnosed a skin disease, if a  picture image was also captured (by doctor or patient) and contributed to a common database, the volume of images would far outpace the current offerings  This gap represents the data exhaust of the clinical encounter. There is valuable information captured in the eye of the professional, used to make decisions and move forward with treatment. The utility of the information carries forward in the mind’s eye of the individual doctor which made the decision about the diagnosis. However, when it is not shared, the utility of that encounter has diminishing returns, as the doctor retires from practice for example.

This is occuring in the context of 50% of the U.S. cellular phone market with smartphones as of early 2012. That means a portion of each of these doctor’s visits (i the United States) happens with a either a patient or physician, or both, with an internet connected camera sitting in their pocket throughout the entire experience. My personal experience with patients and my own family suggests we are utilizing this technology to take pictures of things that happen to our skin. Unfortunately, there is not yet a platform for citizens to participate in the science. 

Thinking forward, this idea is inherently a global project. As robust mobile telephony spreads there is global potential. A chart aggregated on Wikipedia suggests 5 billion people live in countries with >70% mobile phone penetrance.

Allowing the non-professional individual to contribute to the commons is the conceptual next step. The crowd-sourced dermatology database is one platform for one disease category that will attempt to capture that potential.

Much in the same spirit that built open source soft-ware, this data is inherently part of the global commons. Human biology is unique to the individual, but also shared and common to us all. The data from personal pictures of a patient's skin is inherently private-in the sense that it is connected to one individual and part of their personal health record, protected and disclosed at their discretion. However, in another very real sense the image skin forms part of a common knowledge. The coders who produced much of the Linux operating system, which forms infrastructure that supports much of the internet, contributed their knowledge to a common project. It allowed those that came after them to use prior leaps in knowledge to step even further. This pursuit of open knowledge, rich in other fields, has found limited voice in the house of medicine up to this point. This project represents a new kind of trial, a proof of concept. 

Email with questions.

No comments:

Post a Comment