De-siloing Data to Help Improve the Lives of those Suffering from Mental Illness

Thresholds | location Chicago, IL

Objectives

  • Develop and provide Thresholds with a data warehouse to house, scale and process raw data from various sources including Thresholds internal systems, the Illinois Department of Healthcare and Family Services, the Cook County Jail and other potential future data sources.
  • Automate the process of data flow from a central database and create tools to help make managing patient data, generating reports and dashboards easier and more efficient for Thresholds staff, allowing them to better serve their patients.
  • Explore the possibilities of using predictive analytics to help Thresholds better tailor care to people suffering from mental illness.

Findings

  • The DataCorps team built a centralized data warehouse with reporting and analytics tools for staff.
  • The team automated the process of generating reports and dashboards, which had been a very labor intensive process for staff.
  • Several predictive models were tested. While there were inconclusive results, the team was able to develop a foundation of predictive analytics for risk scoring among patients. Thresholds may explore this further once they have fully implemented the data warehouse and reporting functionalities.

Question

Mental health and pressing mental health related issues like substance abuse, particularly opioid abuse, and post-traumatic stress disorder, continue to be serious public health problems in the U.S. Access to quality mental health and substance use services is crucial to addressing these issues and reducing costly hospitalizations, emergency department use, and criminal justice involvement. States cut close to $5 billion dollars from mental health funding during the recession, cuts which came on the heels of years of earlier funding deficits. As a result, many individuals with a mental illness cannot find facilities which offer care and many treatment providers struggle to offer quality care.

Thresholds, a Chicago nonprofit, provides healthcare, housing, and hope for thousands of persons suffering from mental illnesses in Illinois. With 90 community outreach teams, the organization has provided more than 10,000 people with serious behavioral health issues much needed support including illness management assistance, linkage to benefits, housing, employment, talk therapy, medication management, and more.

One of Thresholds’ challenges has been to accurately identify frequently hospitalized and high-risk patients that would benefit from early intervention. Although the data to classify such patients is available, it is often siloed across different agencies and medical centers. Without access to all of a patient’s vitals, it can be difficult for decision makers to optimize and provide proper care and assistance to those struggling with mental illness. Access to such data, when shared and aggregated accordingly, could contribute significantly to Threshold’s knowledge of mental illnesses, help improve diagnostics, and allow staff to deliver tailored and efficiently administered treatment to those in need.

What Happened

In a project sponsored by Bloomberg, four expert volunteers worked on the Thresholds DataCorps project, including Data Ambassador Iakov Gorelik, Project Manager Brennan Lodge and Data Scientists Lulu Cheng and Radu Stancut. The team was successfully able to provide Thresholds with several proof of concept solutions and recommendations for a data warehouse, dashboard and predictive analytics that could be further implemented and improved by Thresholds staff.

To start, the team created a data warehouse comprised of raw data pulled from Thresholds’ internal databases, the Illinois Department of Healthcare and Family Services, and the Cook County Jail. They then developed an automated reporting system, saving staff time and resources, and created a dashboard that staff could use to gain further insights on patients and provide improved quality care for their members; using predictive analytics to help anticipate needs. The new Thresholds dashboard enables staff to visualize data in a more digestible format, identify new trends, and filter among specific patients to better tailor care and allocate resources to reach those most in need.

The team also worked to pinpoint indicators in patient data that would allow Thresholds staff to be more proactive in a patient’s care and intervene earlier to reduce future risk and hospitalizations. To do so, they drew patient information from multiple anonymized data streams on roughly 4,000 patients that were treated by Thresholds from 2013-2015. The rich data streams sourced from the Illinois Department of Healthcare and Family Services, Thresholds and the Cook County Jail provided vital information on patients including case management files, health data, and the treatment and processing of cognitive, wellness, and recovery services. Leveraging the available data, a foundation of predictive analytics for risk scoring among patients was developed.

Next Steps

The work provided by the DataCorps team proved to be invaluable to Thresholds and was also a successful demonstration of the power of collaboration across social service and health sectors and the positive impact it can have. The organization is currently rolling out the automated reports system agency-wide and plans to add more information streams to the new data warehouse. Staff will be able to use the new systems to better tailor care for their members, identify high risk groups, fill care gaps and intervene early. Also, once the buildout of the data warehouse solution, auditing, restructuring and reporting functionality are instituted, there will be even more opportunities for analytics and prediction. Thresholds is exploring further projects with DataKind to continue the work, identifying additional ways to reduce health care system costs and further improve the lives of people struggling with mental illness. In recognition of their work, the DataCorps team was honored by Thresholds at the organization’s annual conference with the “Hero” award.

Scroll to Top