Under every dark cloud, there is a ray of hope. Often, a crisis brings out the best in people and exposes new ways of working and collaborating.

On March 10, Children’s Hospital of Philadelphia, one of the world’s premier healthcare providers for children, went into emergency mode to deal with the rapidly spreading coronavirus. It formed a COVID-19 Command Center that operates 24x7 and consists of representatives from nearly every department in the hospital. Although the hospital has had just 40 confirmed COVID-19 patients so far—fewer than hospitals in neighboring states—it’s taking no chances.

To ensure its readiness to deal with the crisis, the Command Center relies heavily on CHOP’s data and analytics (DnA) team for information about the rate of infection in the regional health system and every aspect of the hospital’s response to it. CHOP uses near real-time data and information to decide how to marshal its resources to contain the pandemic.

Data Analytics in Action

Shakeeb Akhter runs the 80-person DnA team at CHOP. On March 11, he formed an analytics task force to coordinate requests for information coming from all parts of the hospital and serve as a central point of communications. The task force consists of DnA analysts and their counterparts in the business. The group meets daily for 60-minutes to provide updates on work underway, clarify existing requirements, and identify new requests.

Within two weeks, a team of about 20 DnA analysts had built six applications that track and project everything from respiratory diagnoses and Personal Protection Equipment (PPE) to lab tests, telehealth visits, and travel and exposure events. (See figure 1 below for an example.) By mid-April the team had built 18 applications, and the number is still growing. “It’s unbelievable how quickly people have come together to deliver these applications,” says Akhter. “It’s fostered a culture of data-driven decision making that we will continue to strengthen.

Figure 1. Respiratory Diagnoses by Department and Date

About 20 DnA analysts have worked closely with clinician counterparts to build the COVID-19 applications. “A big win for us has been the tight collaboration between business and clinical folks across the enterprise,” says Akhter.

For instance, the analytics task force teamed up with the senior director of supply chain who wanted to track PPE data on an hourly basis. “To build that application, we needed his knowledge about how to classify masks, the different types of masks, the amount of PPE inventory and donations, and how to calculate days-on-hand inventory thresholds,” says Akhter. (See figure 2 for the top-level COVID-19 dashboard that was delivered.)

Figure 2. COVID-19 Supply Chain Dashboard

CHOP’s infection prevention team used the supply chain data to decide whether to require all staff and patients to wear masks at all times and in all facilities. Since there is a nationwide shortage of masks, the team debated the efficacy of a “universal masking” strategy. “We calculated the consumption rate of masks based on badge security data applied to inventory levels over time and determined that we could provide a mask to every person for the foreseeable future,” says Akhter. “That was a big win.”

Similarly, the Provider-Patient Interaction application uses advanced analytics to trace patients and providers who have been exposed to someone in the hospital who has tested positive for the coronavirus. CHOP is also using geospatial data that plots respiratory infections by location to determine where to place its next “drive-through” testing locations. (See figure 3.) CHOP was the first hospital in Pennsylvania to create a COVID-19 test.

Figure 3. Respiratory Diagnoses by Location

Analysts and Infrastructure 

The analytics task force publishes links to the applications on the company’s intranet site. The links provide single sign-on directly into the applications, most of which are built with Qlik’s Qlik Sense self-service analytics product. The CHOP DnA team empowers and trains others in the workforce to develop their own applications. This has resulted in an increase in efficiency with groups, such as Laboratory, to service their own data needs and deploy analytics products on demand.

To ensure data integrity and alignment across the organization, the DnA team developed a COVID-19 ‘stack’ in the CHOP data warehouse. This data infrastructure (i.e., data marts, ETL, and scheduling) offers a single source of truth for all COVID reporting. The stack enforces data definitions across reports and applications and accelerates time to market for new applications. The DnA team works with decentralized analysts to migrate existing reports to the standard COVID-19 data environment.

Most of the DnA analysts involved in COVID-19 projects are members of the hospital’s Quality Improvement team, which helps clinical departments apply data to reengineer processes to improve patient safety, satisfaction, and outcomes and ensure high-quality electronic data capture upstream. Since QI work has slowed with the outbreak of the coronavirus, these analysts have had extra time to devote to the COVID-19 projects. However, DnA’s data engineering and analytics teams have also stepped up to provide critical assistance.

COVID-19 Dashboard

The culmination of all of this work has been an enterprise COVID-19 dashboard that is distributed to enterprise leadership daily and contains data related to Patient Admissions, including those COVID related, Patient Test Results, (daily and cumulative), Outpatient Volumes (including Telehealth Visits), metrics to track the health of our workforce, and critical information related to the supply chain.

Akhter is proud how his team has responded to the crisis and helped CHOP work quickly and effectively to deal with the situation. “We are most proud of our collaboration and speed of delivery, and the impact our applications have had on the organization,” says Akhter. “Our business partners, clinicians, and staff have truly appreciated the value of the team, and have repeatedly recognized that the insights we’ve provided have allowed them to make better, data-informed decisions during a time of crisis.”