FAQs

General FAQs

WHERE CAN I VIEW THE APPLICATION?

You may download a copy of the application here. Applications must be submitted on the online platform to be considered eligible. Emailed applications will not be accepted.

WHERE DO I UPLOAD MY SUBMISSION?

Where can I find the slides and recording of the April 18th Webinar?

can i compete in any stage of the ai health outcomes challenge directly?

The AI Health Outcomes Challenge is structured in such a way that you may not compete in Stage 1 or 2 directly without participating during the Launch Stage and being selected to move on to subsequent stages.

CAN participants (individuals, teams or organizations) MAXIMIZE ODDS FOR WINNING BY PROVIDING MULTIPLE SUBMISSIONS?

Multiple submissions are allowed, but each one must be a unique approach, sufficiently differentiated, and in compliance with the Challenge requirements.  Although you may have a greater likelihood of selection, only one submission will be selected.

Are there word limits on the questions in the launch stage online application?

Yes there will be word limits. We will be providing further guidance on the expected length of application responses. Please make sure you have registered at ai.cms.gov or check the website frequently for updates.

Is there a way to subscribe to updates to the challenge?

We encourage individuals to register for the challenge at ai.cms.gov to receive updates about the challenge.

Data FAQs

WHAT SPECIFIC CMS DATA SETS WILL BE PROVIDED TO PARTICIPANTS IN STAGE I?

Stage 1 participants will receive a random, continuous 5% sample of Medicare fee for service (Part A and Part B) claims data. The initial data file will encompass approximately 3 million beneficiaries over a 5-year time frame. This includes home health, hospice, inpatient, outpatient, skilled nursing facility, carrier standard, and durable medical equipment standard analytical files. Part D claims data is not a part of the LDS.

You can find CMS’ data dictionaries here.   

What Medicare Parts are included in the data file?

Stage 1 participants will receive a random, continuous 5% sample of Medicare fee for service (Part A and Part B) claims data. Part C and Part D data are not be included.

IS THE LIMITED DATA SET (LDS) STRUCTURED AT THE PATIENT LEVEL?

Yes. The datasets contain beneficiary-level health information but exclude specified direct identifiers. The LDS (limited data set) is comprised of standard analytical files based on institutional and non-institutional claim files.

IS THE LIMITED DATA SET (LDS) A LABELLED DATASET (OUTCOMES ALREADY KNOWN)?

Participants who are selected for Stage 1 of the competition will request and receive an LDS with 5 consecutive years of Medicare beneficiary data, encompassing around 2.5-3 million unique beneficiaries. Participants who are selected for Stage 2 of the competition may request an additional 5 consecutive years of data from the same beneficiary sample.

The LDS are structured, labeled datasets where health outcomes are known. Please refer to the CMS’ data dictionaries to learn more about the LDS fields, labels, and data types to craft your proposal.

WILL THE LIMITED DATA SET (LDS) BE MADE AVAILABLE TO PARTICIPANTS DURING THE LAUNCH STAGE?

No. In the Launch Stage participants are only being asked to describe their concept and, as such, the LDS is not needed. Please refer to CMS’ data dictionaries to learn more about the LDS fields, labels, and data types to craft your proposal.

If Launch Stage participants apply to purchase an additional LDS for their own use, prior to their selection as a Stage 1 participant, CMS will not reimburse the costs of the additional LDS. 

CAN WE RECEIVE A SAMPLE OF THE LIMITED DATASET IN THE LAUNCH STAGE?

No. Please refer to the CMS’ data dictionaries to learn more about the LDS fields, labels, and data types.

WILL THE STAGE 1 AND STAGE 2 LIMITED DATA SETS (LDS) BE FREE?

The LDS will be provided free of charge to the Stage 1 and 2 participants deemed qualified to receive it. Participants will be required to sign and comply with a Data Use Agreement (DUA) that includes specific rules and requirements for receiving, storing and protecting the data.

If Launch Stage participants apply to purchase an additional LDS for their own use, prior to their selection as a Stage 1 participant, CMS will not reimburse the costs of the additional LDS. 

The LDS are structured, labeled datasets where health outcomes are known. Please refer to the CMS’ data dictionaries to learn more about the LDS fields, labels, and data types to craft your proposal.

ARE INDIVIDUALS WITH MEDICAID INCLUDED IN THE LIMITED DATA SET (LDS)?

The data that will be provided for the Challenge is the Limited Data Set (LDS) which contain beneficiary-level health information, but exclude specified direct identifiers as outlined in the Health Insurance Portability and Accountability Act (HIPAA). While the file may include individuals dually eligible for both Medicare and Medicaid, the primary focus of the data — and the Challenge — are Medicare fee-for-service beneficiaries.

IS FREE TEXT DATA INCLUDED IN THE LIMITED DATA SET?

We understand the value of processing free text data; however it is not currently available in the Limited Data Set (LDS). We are open to creative ways to incorporate such unstructured data into the AI solution.

CAN DATA SOURCES OTHER THAN THE LIMITED DATA SET (LDS) BE USED FOR THE CHALLENGE?

Participants may use de-identified, public data sets to the LDS (e.g., Census data). Participants are not required to use other de-identified public data, however, they will be evaluated on their ability to identify other data sets and types of information that will be useful to further refine their solutions following the competition. Participants must follow the Data Use Agreements for the LDS.

ARE PROPRIETARY DATA SETS ALLOWED IN THE CHALLENGE?

The use of proprietary data is prohibited at the Launch Stage. It is critical the competition is fair and merit-based with access to non-publicly accessible data giving certain participants unfair competitive advantage. 

If an EHR company is participating, can they use the data they already have?

No, to maintain fairness, only publicly available data is allowed at the Launch Phase of the challenge.  

Will participants who receive the LDS be able to use the data for research outside -- or after -- the challenge?

No. participants will only be allowed to use the LDS for the purposes of the challenge and will be expected to destroy the files and submit a Certificate of Disposition (COD) form to close their DUA once the competition is complete.

Eligibility FAQs

Who can participate in the AI Health Outcomes Challenge?

Anyone can apply to participate in the challenge, although by law, prize money can only be awarded to citizens or permanent resident of the united states or private entities that are incorporated in and maintain a primary place of business in the united states.

CAN NON-U.S. CITIZENS PARTICIPATE IN THE ARTIFICIAL INTELLIGENCE HEALTH OUTCOMES CHALLENGE?

Non-U.S. citizens and nonpermanent residents can participate as individuals or as a member of a team that otherwise satisfies the eligibility criteria, but these individuals will not be eligible to win a monetary prize (in whole or in part). However, their participation as part of a winning team, if applicable, may be recognized when results are announced. However, provided that a participant team contains at least one U.S. citizen or permanent resident/US incorporated entity, that team can be awarded prize money. Specifically, the prize money in such a scenario will be given to the U.S. citizen/entity.

IP FAQs

DO I RETAIN INTELLECTUAL PROPERTY OWNERSHIP?

Yes, each entrant retains full ownership and title in and to their submission. As an entrant, you reserve all intellectual property rights not expressly granted under the challenge agreement.

However to receive an award, participants must grant to the Federal government the nonexclusive, nontransferable, irrevocable, paid up license to practice, or have practiced for or on its behalf, the solution throughout the world for Federal government purposes. Each participant also warrants that the work is free of security threats and/or malware.

Are winners required to provide ongoing maintenance and support for solutions for free after the Challenge ends?

Participants are not expected to provide ongoing maintenance and support for any solutions developed for the AI Health Outcomes Challenge.

The rules related to IP rights apply to all entries, regardless of whether they are chosen as the grand prize winner and whether participants accept the prize money.

Do the Intellectual Property (IP) rules apply if prize money is not accepted?

The rules related to IP rights apply to all entries, regardless of whether they are chosen as the grand prize winner and whether participants accept the prize money.

CAN PARTICIPANTS WORK AND FINDINGS BE PUBLISHED IN SCIENTIFIC JOURNALS?

Yes, winners will retain ownership of their intellectual property rights in the solution and therefore can publish in scientific journals; however, the Federal government also retains the right to the solution.  Please also refer to the discussion regarding “research” in the public announcement.

JUDGING FAQs

What is the background of the judging panel?

The judging panel will consist of a mix of data scientists, technologists, clinicians, and patients/advocates.  Applications will be reviewed and scored by at least three judges from varying fields. 

How will Stage 1 and Stage 2 models be evaluated?

The Stage 1 and Stage 2 judging criteria are being refined and more information will be provided at a later date.

TEAM FAQs

WHAT SIZE SHOULD A TEAM BE? IS THERE A MINIMUM, A MAXIMUM, OR AN IDEAL NUMBER OF MEMBERS PER TEAM?

There are no requirements or recommendations for team size, but cross-functional teams are encouraged. It is up to participants to build their teams, determine roles and, if applicable, split up any prize amounts.

IF THERE ARE INDIVIDUALS WHO ARE INTERESTED IN DEVELOPING TEAMS FROM OTHERS WHO HAVE EXPRESSED INTEREST, CAN CMS PUT THESE INDIVIDUALS IN TOUCH WITH EACH OTHER?

CMS does not currently have a mechanism to help individuals connect with potential team members.  We recommend using social media (i.e., Twitter), using the hashtag #cmsaichallenge to connect with other participants.

Page Last Update: May 17, 2019

Close Menu