About: In May 2013 the US government released a bunch of data on how much hospitals charge for various procedures - see my comments here. This is the first time we've had such extensive data on hospital prices, and hopefully more transparency will lead to more reasonable prices.
Please - write some good apps with this data and tell me about it!
Database info: Here are the tables in the database:
hospitals- name and address of each hospital, as well are their latitude and longitude. The dataset contains 3337 hospitals.
discharges- a description of each discharge (i.e. procedure) that we have data on. These are the 100 most common inpatient procedures done by hospitals that are charged to Medicare, and the 30 most common outpatient procedures.
prices- a hospital, discharge, and the average covered charges and total payments for that discharge at that hospital, as well as how many times that happened.
averagePrices- a discharge and the average covered charges and total payments across all hospitals.
hospitalPriceIndices- a hospital and their price index with respect to covered charges and total payments. I defined the price index to be how their prices relate to the national average - a price index of 100 means exactly average, 150 means 50% more expensive, and 50 means 50% cheaper.
totalPerformanceScores- a hospital and its clinical process score, patient experience score, and total score. All are out of 100, and higher is better. See the original data source for more information.
Covered charges vs. total payments: I am not a health care pricing expert, but from the media coverage it sounds like "covered charges" is what the hospital bills you if you don't have insurance, and "total payments" is what Medicare pays. Most stories have focused on "covered charges".
Using SQLite on Windows Phone: I followed this handy guide and was up and running in no time!
How I made the database: I wrote a few Python scripts. generateSqlite.py generates the database and calculates the average prices and price indices. It also generates a file that I passed to bingmapsportal.com for geocoding (by using doGeocode.py). I manually saved the results to a file, then ran generateSqlite.py again to pull in the latitude/longitude.