Here is a preview of the RocHackHealth 2016 challenges! Full information and data will be posted on Friday April 8, 2016. But just in case you are curious:
- Challenge #1: Predicting Hospital Re-Admissions: In this challenge, your team will try to predict which patients will be re-admitted to the hospital after being discharged from a hospital stay. This is a real problem for hospitals, who don’t get paid for a re-admission if it happens <30 days after the patient was discharged. There are a million companies trying to develop algorithms to do this, so here is your chance to show your stuff. You will be given a training data set of several files, and will need to assemble them and train your algorithm. You are free to use any algorithm you wish that your team codes, or combine several. On Sunday at 9:00 AM, we will release the validation data set (which does not tell you who got re-admitted) and you will run your algorithm and submit your predictions. The team with the most accurate predictions, lowest false-negative rate, best presentation of their results and cuspy methods, and best dark magic wins $100.
- Challenge #2: Follow the Narcotics: It’s a cross between CSI Medicare and Breaking Bad. In this competition, teams will look for unusual patterns of narcotic prescribing in Medicare data. Medicare is the US version of national healthcare, and provides health insurance for Americans 65 and older who have paid taxes into the system when they were employed. It also covers people <65 years old with disabilities, kidney disease who need dialysis or have a transplant, and who have the neurologic disease amyotrophic lateral sclerosis. You will be using a version of a publicly available file that shows which providers (doctors, nurses and healthcare providers) prescribed which medications during 2013 that were paid for by Medicare. You will use the clustering and pattern recognition methods of your choosing to identify providers who prescribe lots of narcotics (morphine, dilaudid, oxycodone, etc.), and identify their features (state, geographic location, specialty). To accomplish this challenge, you will also need to display your results using amazing graphics. The team with the most elegant identification methods and zero-day visualizations will win a $100 prize.
- Challenge #3: DocGraph Provider Community Identification: Social Network Analysis meets Healthcare in this challenge. Dedicated to Fred Trotter, the original healthcare hactivist, in this competition teams will use clustering methods and network science methods to identify “teams” of Medicare providers linked to each other by shared patients in the state of New York. It’s really SNA of how healthcare providers are linked to each other by shared patients. You will use a subset of a publicly available Medicare dataset that contains only NY providers (doctors, organizations, nurse practitioners, physician’s assistants, etc.). You will need to link this data set to another file that contains information about each provider (specialty, location, etc.), and to another file which contains the geolocation coordinates of the provider at the zip code level. You should identify closely linked teams of providers, and figure out their characteristics (team composition, etc.) and any other interesting features of how they are connected. To accomplish this challenge, you will also need to display your results using innovative graphics. The team with the most elegant methods and visualizations will win a $100 prize.
We hope to see you there!