Research IT launches phase I of STARR-Radio, a new platform ability to bring Radiology data to Stanford researchers

Research IT has built a cloud scale radiology imaging repository that brings data from Hospital picture archiving and communication system (PACS) into a research archive.  Stanford has half a petabyte of historical radiology data and ~20% of this data is now in the research archive as part of phase I and we continue to add the rest. The data contains different modalties e.g. X-rays, MRI, CT, Ultrasound, Mammography, etc. Research IT has developed robust and scalable de-ID pipelines for X-rays, MRI and CT as part of phase I release and is continuing to build out for rest of the modalities. Researchers can request Research Informatics Center for access to the imaging data. 

STARR-Radio is being developed in collaboration with Research IT faculty committee and Google Healthcare team.

PCORI study launched to determine whether cognitive behavioral therapy or pain self-management classes are effective at alleviating pain and reducing opioid use among people with chronic pain

EMPOWER, a large PCORI grant funded multi-site trial, under leadership of PI Beth Darnall, was launched in early July 2018. Nearly  ~1300 participants from Stanford Pain Clinic, Stanford Primary Care, Intermountain Health Primary Care, Stieg Pain Clinic and Phoenix VA Primary Care will participate in the research study. This research involves a 3-arm Randomized Controlled Trial comparative effectiveness trial of two evidence-based behavioral treatments, cognitive behavioral therapy for pain and chronic pain self-management, within the context of voluntary, patient-centered opioid tapering. This research builds on Darnall team's first report on "Patient-Centered Prescription Opioid Tapering in Community Outpatients with Chronic Pain".

Research IT team (Teresa Pracht, David Tom and Garrick Olson) deployed an instance of HIPAA compliant CHOIR, customized for the PCORI study, on Google Cloud Platform, to support the multi-site collaborative research and analysis. The study team is using Research IT managed Stanford REDCap for storing study protocols.

Palliatve care prediction model developed by Stanford lab is deployed at the Stanford Hospital

Late last year, MIT press review published on a Stanford developed deep learning model for palliative care. Since then, the Shah lab, co-authors of the paper, have developed multiple other models.There is now a gradient boosted model (GBM) and a brand new probabilistic model. The GBM model was deployed at the Stanford Hospital in Jun 2018 (link).

Research IT team member, Tina Seto, worked collaboratively with Nigam Shah's lab and Stanford Health Care to develop the data pipeline that fed the models. First version of the model was developed using data in STARR datalake and the production model was developed using SHC Clarity data to reduce the data latency from 3 days to 1 day.

With Stanford Center for Population Health Science

Under a new DASHER initiative, Research IT's Susan Weber has made a subset of de-identified STARR structured data (STARR-Tahoe) available on PHS Data Center portal. This allows for new way of exploring STARR data that did not exist before. In particular, you can quickly navigate through the metadata and table descriptions, and visualize the distribution of data associated with the variables. STARR-Tahoe will be updated at least annually. 

Stanford Center for Population Health Science (PHS), under the direction of Mark Cullen, Senior Associate Dean for Research, focuses on study of health determinants and outcomes in large populations. It sits at the intersection of medicine and public health, spanning basic and social sciences, enabling integrated research that encompasses virtually every domain of life and society. The Data Center at PHS hosts a diverse collection of high-value “marquee” data sets like Truven Marketscan and partners with several data owners around the world like Born in Bradford. 

With Stanford Technology Analytics and Genomics in Sleep (STAGES) Program

Alliance Sleep Questionnaire (ASQ) Stanford Technology Analytics and Genomics in Sleep (STAGES) enrolled their first patient at Stanford Sleep Clinic in April 2018. Remaining 10 sites will go live soon. This is a significant expansion to the existing ASQ program. ASQ has been engineered by Sanjay Malunjkar at Research IT in collaboration with Eileen Leary, Sr. Manager for Clinical Research at the Sleep Clinic. ASQ first went live in Stanford Sleep Clinic in 2011 and has since enrolled 8,000 patients. The STAGES program will expand ASQ to 30,000 patients.

Here are some additional details on the STAGES research program from Eileen. The STAGES study is funded by the Klarman Family Foundation and will collect multi-modal data on 30,000 patients across 11 different sleep clinic sites (across USA and Canada). Sleep disorders and sleep dysregulation impact over 100 million Americans, causing enormous medical consequences including cardiovascular (arrhythmia, hypertension, stroke), metabolic (diabetes, obesity) and psychiatric (depression, irritability, addictive behaviors), to name only a few. Sleepiness costs millions due to workforce errors and accidents. Although the global market for just hypnotic products will reach $76.7 billion by 2019, our fundamental understanding of sleep remains a biological “black box”. This curtails our ability to identify etiologies and thus treat sleep disorders effectively. To address this unmet need, STAGES team proposes to conduct large-scale genomics studies, sleep phenotyping and automated PSG data analysis. The information will be crucial for our understanding of the genetic architecture of sleep and to improve detection, treatment and prevention of sleep disorders. To maximize research potential, all tools and data will be made available to the research community. Specifically, Stanford is responsible for collecting data for 30,000 individuals from multiple sites:

• On-line sleep/medical history questionnaire (the ASQ, a branching logic questionnaire developed by Harvard University, University of Pennsylvania, University of Wisconsin-Madison, St Luke's Hospital, and Stanford University)
• In lab nocturnal Polysomnography data (one night PSG)
• Computerized Neurocognitive Battery (U Penn CNB)
• Actigraphy over 2-4 weeks (Amazfit Arc)
• 3-D craniofacial images.
• Stored biological samples (DNA, plasma, serum) for future biomarker research
• Genome Wide Association Study (GWAS) data

This study will be a landmark that will constitute a gold standard for the analytics of sleep. In collaboration with Prometheus Research, the STAGES program has built a comprehensive data platform to allow for easy data collection and management. 

With CHOIR Learning Healthcare Platform

Collaborative Health Outcomes Information Registry (CHOIR) team (Garrick Olson, Teresa Pacht, David Tom, and ex-team member Randy Strauss), working in close collaboration with Prof Sean Mackey, MD, PhD, Chief of the Division of Pain Medicine, delivered the point-of-care clinical trial feature on the CHOIR platform. This method was developed by Stanford University biostatistician Philip Lavori, PhD, and researchers from VA Boston Healthcare System in 2011 as an alternative to expensive, lengthy, double-blind, placebo-controlled clinical trials to compare drugs and procedures that are already in regular use. A layperson summary of the method can be found here and the publication is here.

Research IT is supporting a prestigious new $8.8M PCORI grant, led by Beth Darnall, PhD Clinical Professor of Anesthesiology, that will use a point-of-care clinical trial in patient-centered opioid tapering with active behavioral treatment. The study will include ~1000 participants from Stanford Pain Clinic, Stanford Primary Care, Intermountain Health Primary Care, Stieg Pain Clinic and Phoenix VA Primary Care.