The Streetlights Project brings the clinical data shared with the Chicago Health Atlas and HealtheRx projects together as a shared and dynamic data resource for the public, and public health officials, independent of the healthcare system.
We’ll extend our work to create dynamic maps of chronic disease burden, and control. For example we may map the average percent Hemoglobin A1C (a laboratory measure of blood sugar levels) or average blood pressures as measured by local healthcare institutions by ZIP code. Patients may interact with our website to self-identify as a diabetic within a ZIP code to see how they compare against the average Hemoglobin A1C for their ZIP code and then be presented with the most relevant local resources to help them manage their condition. Public health officials can identify neighborhoods with high disease burden, perform targeted interventions, and monitor progress as measured by health care providers contributing data to our system.
We’ll specifically expand the number of participating healthcare institutions and improve the speed by which our system updates clinical data. We’ll also extend our community resources database to include all of the City of Chicago and actively engage community residents on how to use Streetlights and incorporate their feedback to continuously update and improve the data on our site.
Refinement Phase entry below:
Our project team, their contributing expertise and the partnering agencies they represent reflect the major stakeholders and roles defining Streetlights: Abel Kho (Northwestern University, Project Streetlights lead and faculty in General Medicine and Biomedical Informatics), Katie Jackson (Northwestern University, data analyst), and Satyender Goel (Northwestern University, project manager); Stacy Lindau and Karen Lee (University of Chicago, leading the South Side Health and Vitality Studies), Eric Jones (Chicago Department of Public Health, epidemiology and data realist), Brad Malin (Vanderbilt University, Health Information Privacy expert and project consultant), Jorn Boehnke, John Eric Humphries and Scott Kominers (University of Chicago and Harvard University, algorithm development and economic analysis) and Fred Rachman (Alliance of Chicago Community Health Services, engagement of Community based primary care providers, patients and resources).
Clinical data and resources have been contributed by Chicago based institutions including: Northwestern University, Rush University, University of Chicago, University of Illinois, Cook County Health and Hospital Systems, the Alliance of Chicago, and most recently Loyola University. Initial support for the effort has come from the Otho Sprague Memorial Institute (seed funding), the Chicago Community Trust’s Smart Chicago Collaborative (Dan O’Neil, Chicago Health Atlas website development at www.chicagohealthatlas.org and project management), the Chicago Department of Public Health (Advocacy, public support, and convening), and the University of Illinois Masters in Information Sciences program (student intern teams).
* How do you know there is demand for this project?
Much of our initial work was accomplished from a variety of funding sources (foundations, federal, state), brought together because of the shared belief in the importance and effectiveness of pooled data to tell the story of health and guide future public health interventions. For years, we have had data sharing agreements and trusted public health/research relationships in place with the major healthcare institutions in Chicago which have enabled merging of de-identified records for over 5,000,000 unique individuals. We have been approved to use our software to link Veteran’s Affairs (VA) data and non-VA health institution data to help identify how often veterans seek care outside of the VA system, and the NIH has granted similar approval to link NIH-sponsored cohort data to our Electronic Health Record (EHR) data in order to supplement and hopefully improve the usefulness of existing cohort studies. We have been invited to present our approach to academic partners, public health associations, and community foundations.
* How is your project different from what already exists?
This project is innovative in its technology approach, how the approach enables contribution of data more openly while protecting privacy, and the the inclusion of meaningful community environmental and resource data. …Unlike current approaches to pooling clinical data across sites that require either extensive agreements to enable sharing of identified data, or share only aggregated de-identified counts of data, we have developed a technology that allows us to focus on aggregate deep and comprehensive clinical data containing as complete a longitudinal health record as possible (given institutional approval and the protection of privacy). . We created a standardized algorithm to assign a unique ID to individuals wherever they seek care; this ID is embedded into a software application to enable linkage at the individual level, while still removing all protected health information (other than zip code to allow for mapping).
This creates a rich database covering a concentrated geography spanning diverse neighborhoods and populations which can then be paired with our inventory of relevant local community resources.
* How will the data or information you use or create be made open?
We have already shared clinical data through the www.chicagohealthatlas.org website, and community data through www.southsidehealth.org and will continue to expand and integrate these data as part of the Streetlights Project. As health care providers, we are keenly aware of the balance between openness, and privacy considerations embodied by the Health Insurance Portability and Accountability Act, and its recent update through this year’s HIPAA Omnibus rule. Much of our work over the past few years has focused on creating a secure software framework that eliminates the risk of either provider level or patient level re-identification so as to enable traditionally competitive organizations to participate in our data integration project. We will make available our data through an API for consumption by sites beyond our current Chicago Health Atlas partnership.
* What will you make or do in this project?
We will (1) refine our software platform to enable more real-time uploads of de-identified data for rapid visualization, (2) expand the scope and depth of our community resource database, and (3) develop additional functionality to enable individuals to interact more easily and contribute additional data to the Streetlights Project database. The goal is a plug and share client side application which can enable peer-to-peer sharing (instead of our current centralized aggregation system) of clinical data with tight access and security controls. Two specific features we will develop will be a stratification and weighting engine (to account for over or under-sampling of specific demographics within a ZIP code or neighborhood), and a distributed query tool (initially will be an extension of existing open source software). Our disease focus will be driven by our stakeholders, and to date has included chronic diseases such as diabetes, asthma, and hypertension, but in the future we may include other considerations such as wellness, care fragmentation, and crime/violence.
* How can others learn from/build on what you do?
Our data platform enables linking in of data at the individual level and is industry agnostic. In addition to the uses noted above (VA, NIH), individual sites have used our software to identify deceased patients by linking to National Death Index data, or to remove duplicate records from within their databases. We anticipate that other industries may find the ability to assign a common identity to individuals to be highly valuable for data aggregation and analytics. Community stakeholders or innovators may choose to use our software to integrate in additional data into websites, including community calendars of health-related events or up to date data on environmental contaminants (e.g pollen counts/air pollution index).
* How much do you think it will cost? For the roles below we estimate approximately $750 thousand per year over two years, adjusted based on allowed ranges for fringe/indirects. As appropriate, we are prepared to submit a more detailed budget and justification to factor in exact salaries/fringe and timeframe for deliverables.
* How would you use News Challenge funds? To enable dedicated effort to rapidly expand our project scope. We anticipate applying funding to the following specific roles: Streetlights Project lead (50%, project oversight, design, promotion), web developer (100% effort, API development to connect with public websites, visualizations development), software developer(s) (Java and .Net at 50% each for integration of existing software, build out of new functionality), data analyst (100%, develop recurrent automated data extraction processes), data architect (25%, develop common data model), health information privacy consultant (20%, HIPAA and Omnibus compliance audits), epidemiologist (50%, development of stratification/weighting engine rules), and project administrator time/effort (25%, grants administration, institutional agreements or approvals as necessary), Community resource mapping project lead (100%, Community partners engagement, MapsCorps training and placement, and data curation), MapsCorps interns (5 x 100%, ground-truth mapping).
30 second video link: