If you have a “Hospital Account” (i.e. @mountsinai.org) you can request a school account here
Access requests will be submitted via SailPoint – to access Sailpoint you will need to be on the Mount Sinai Network (i.e. on-site and connected to WIFI) or connected to Mount Sinai VPN.
To request access to datasets with personal health information (PHI) (i.e. all datasets but Deidentified OMOP-mapped EHR data and DE-SynPUF OMOP), you will need an approved IRB referencing AIR·MS platform (IRB # 20-01288) to access and store data.
This IRB will need to be attached to the DUA form during access request on SailPoint – please make sure to have this in place before requesting AIR·MS access for PHI-containing datasets.

Q: How do I request access to AIR·MS datasets?

A: Once you have all requirements (see above) – you can request access to AIR·MS via Mount Sinai SailPoint. You can find a detailed guide on access request here.

On SailPoint you will search for appropriate access options:

Dataset	Mount Sinai User Selection	HPI Potsdam Users Selection	Approval Process
DE-SynPUF data (synthetic OMOP dataset)	AIR·MS Production DE-SynPUF OMOP (MSSM)	AIR·MS Production DE-SynPUF OMOP (HPI)	Automatic approval after submission
De-identified OMOP-mapped EHR data	AIR·MS Production OMOP De-ID (MSSM)	AIR·MS Production OMOP De-ID (HPI)	Automatic approval after submission and DUA signature on SailPoint
OMOP-mapped EHR data with PHI	AIR·MS Production OMOP PHI (MSSM)	AIR·MS Production OMOP PHI (HPI)	IRB required and manually approved after review
Pathology Metadata	AIR·MS Production Pathology Metadata (MSSM)	AIR·MS Production Pathology Metadata (HPI)	IRB required and manually approved after review
Radiology Metadata	AIR·MS Production Radiology Metadata (MSSM)	AIR·MS Production Radiology Metadata (HPI)	IRB required and manually approved after review

Abbreviations: AIR·MS, Artificial Intelligence-Ready Mount Sinai; De-ID, De-identified; DUA, Data Use Agreement; EHR, Electronic Health Record; HPI, Hasso Plattner Institute; IRB, Institutional Review Board; MSSM, Icahn School of Medicine at Mount Sinai; OMOP, Observational Medical Outcomes Partnership; PHI, Protected Health Information; SYNPUF, Synthetic Public Use Files.

ATTENTION: remember to use the right access requests if you are a Mount Sinai employee vs a HPI Potsdam user.

Q: I submitted a SailPoint request—how do I know if it’s approved?

A: You will be notified via email once your request is processed. If your request is denied, you will receive details on the grounds for denial and guidance on corrective measures that can be undertaken. The process takes 7-10 business days and is subject to variation.

Q: How do I track the progress of my access request?

A: You may monitor the progress of your request status by logging into SailPoint.

Q: My access was denied. What should I do?

A: Make sure you selected the correct role:

- Mount Sinai users: AIR·MS Production <dataset name> (MSSM)
- HPI users: AIR·MS Production <dataset name> (HPI)
Also make sure your IRB covers usage of AIR·MS explicitly if you requested access to non-deidentified datasets.

GETTING STARTED & ENVIRONMENT SETUP

Q: What are my username and password?

A: To connect to AIR·MS you will need to provide your username and password.

- Your username corresponds to your Sinai Network ID (e.g. “Angela Pattinson” will have a Sinai network id like pattia01).
- Your password is your Sinai password used to connect to any Sinai resources.
You WON’T need to add the VIP token to your Sinai password.

Q: How do I connect to AIR·MS?

A: To connect to AIR·MS you need to be on the Mount Sinai Network.

- Connect to the Mount Sinai network either by connecting to the LAN or MSSM Green while on campus or by using the VPN connection if not on campus.
- If using VPN, make sure you click on the “Tunnel” button on the VPN splash screen.
You can follow the tutorial to connect to AIR·MS from Python or R environments.
- ATTENTION: you need to be on Mount Sinai Network to access the GitHub Page.
You can connect to AIR·MS via your local device or from Minerva High Performance Computing (HPC).
- ATTENTION: to connect from your local device you need to be compliant with cyber-security and install Microsoft Company Portal on your device.

Q: I can’t reach the SailPoint or GitHub tutorial pages.

A: Ensure you’re connected to the Mount Sinai network or VPN.

Q: Can I connect to AIR·MS using a database client?

A: We do not actively support database clients to connect to AIR·MS. However, you can configure clients, for example DBeaver, on your own to connect to AIR·MS. You can find a guide on how to use DBeaver here: Researcher Tutorials.

- ATTENTION: if you connect from your personal device, the device has to be compliant with cyber-security by installing Microsoft Company Portal.

DATA

Q: How is data stored on AIR·MS?

A: AIR·MS uses an extremely fast, in-memory databased called SAP HANA (High-performance ANalytic Appliance) to stored structured datasets. Each dataset is stored in a separate schema on our AIR·MS HANA database. Users need to request access to each schema separately and be approved.

Additional unstructured data (e.g. DICOM images, Pathology Whole Slide Imaging) are stored on Minerva and can be linked to structured AIR·MS data through specimen IDs.

Q: What data are available on AIR·MS?

A: The current list of datasets and their attributes is available on AIR·MS can be found here.

Q: How can I link different datasets?

A: If you have access to multiple schemas at the same time, you can link tables using the common Epic MRN that is stored in PHI-containing datasets.

Q: How do I query AIR·MS?

A: To query data on AIR·MS you can directly use Structured Query Language (SQL), or you can use language-specific (R or Python) libraries that allow you to interact with the database more intuitively. You can find more information here.

More information on SQL and SQL HANA dialect can be found here.

Q: Can I create a dataset on AIR·MS to share with my collaborators?

A: AIR·MS does not support multi-user personal schemas yet. However, every user is provisioned with a user schema named as their Mount Sinai user ID on which they can store personal tables.