Sponsored Content
GE Healthcare and Intel Corp.

The Journey from Business Intelligence to Artificial Intelligence to Advance New Models of Care

How health care organizations use data as keys to a smarter future, and opportunities and challenges

The health care field has been progressing slowly through three phases of data management: data collection, sharing and analytics. Some organizations are only just beginning to have the necessary analytics capabilities that enable quality, operations, financial and clinical improvement. Others are well on their way toward achieving a truly data-driven culture.

How Organizations Are Using Data as They Advance in Their Analytics Journey

A group of hospital executives recently gathered in Chicago at the American Hospital Association Executive Forum on Advancing Affordable Health Care to discuss how they are using data analytics to ensure the delivery of high-quality, affordable, equitable care. Several reported that their use of business intelligence (BI) or artificial intelligence (AI) has resulted in substantially improved patient outcomes and smarter strategic decisions.

Improving Processes and Performance

At Elgin, Ill.-based Advocate Sherman Hospital, part of the Advocate Health Care system, field advisers share data with physicians and office managers at primary care practices to highlight one or two opportunities to improve clinical integration targets. Bruce Hyman, M.D., vice president of medical management, said, “It's looking at every possible way to be able to share data and make a difference.”  For example, the colon cancer-screening rate wasn't on target or hemoglobin A1Cs were not where they needed to be.

Jean Scallon, CEO at behavioral health-focused Bloomington (Ind.) Meadows Hospital, said her hospital uses genomic testing to ensure that patients are on the right medications, and is looking to partner with someone to remotely monitor devices worn by children with attention-deficit hyperactivity disorder and depression to allow for care intervention when it spots high activity or sleep patterns to keep kids out of the hospital and functioning in school.

University Hospital, a Level I trauma center in Newark, N.J., has used data to analyze how its emergency department processes patients, according to John Kastanis, president and CEO. “We saw dramatic improvement in the time for patients to see a doctor or get a bed, and made a real dent in the number who left before being seen.”

A panelist from GE Healthcare’s Applied Intelligence unit, which is embedding AI in medical devices, remarked that even when inundated with data, organizations can “start small and start now” to bring about large improvements. Kastanis agreed, saying that at his organization, “the beauty of AI is we can take specific electronic health record data and apply [them] to productivity and financial performance and get some quick results.”

Driving Services

Many participants noted that they use collected data to drive strategic decisions on the kinds of services they offer and where in the community they provide them.

Andrew Goldfrach, CEO of University Hospital Avon Rehabilitation Hospital in Cleveland, said that his health care system is studying where patients are willing to travel for certain services, and what specialties are appropriate by facility for the area’s population.

Likewise, said Kastanis, “You need data to know when you don’t have the patient volume to sustain transplants and all the resources required.” Clinical-efficiency data analysis has helped both Kastanis’ University Hospital and Temple University Health System in Philadelphia to make decisions on continuing or building their tertiary capabilities. Temple’s director, Bob LeFever, said, “We are one of the leading transplant places in the state for hearts and lungs and it’s, of course, what’s right for the community.”

Predicting Needs

Yandong Liu, assistant professor/Health Disparity Research Fellow at Chicago’s Rush University, said that Rush has conducted analysis, based on existing data on demographics and education levels, as well as Rush’s own annual Community Needs Assessment to determine potential ways to reduce health care disparities in Rush Medical Center served community areas.

Rush University’s Health Systems Management Department has also planned to use newly developed job market predictive software as an additional tool to look at the future of the health care job market and decide which educational programs to develop further to support the workforce roles that will be most in demand.

Challenges Organizations are Encountering on Their Analytics Journey

A Lack of System Interoperability and a Subsequent Need for Greater Data Sharing and Collaboration

Steve Johnson, GE global account director for Intel Corp., asked the group if they thought incompatible technology and data residing in siloes was limiting the sharing of important data among providers.

LeFever said, “This becomes important when it comes to prescriptions. At one of our system’s hospitals, the emergency department’s electronic system is not interoperable with the hospital's electronic system. So the ED staff are trying to determine what the drug usage is of a person coming in and cannot find out what this person’s history has been at the hospital except manually or by phone.”

Hyman despaired that many providers must still fax documents to each other, and then scan them before they enter a patient’s electronic health record (EHR). “The time, the waste, the paper, the machinery!”

While Scallon’s hospital collects data on health outcomes, “it’s all on paper,” she said. Without federal funding for EHR system implementation in behavioral health organizations, few have a way to electronically record and track various patient health data. Furthermore, Community Health Needs Assessment, suicide and opioid use data “are kept in little pockets across the state” and is often old, inhibiting how well problems can be addressed.

A Lack of Data Analytics Expertise and Experience Among Providers

One of the biggest challenges in data analytics utilization is changing providers’ mindsets. No matter how accurate the data on outcomes or local benchmarks or national trends they are shown, Goldfrach said, “Some providers, who have been doing their role for years and years in a certain way, are not necessarily willing to change their thought process” on applying data.

A panelist from GE Healthcare, conversely, expressed that members of the younger generation are “digital natives” who grew up with smartphones. They expect a high level of data analytics and see the potential for unusual applications like gamification to improve staff performance metrics.  

Kastanis believes that most providers don’t have the sophisticated expertise needed to build and use a data warehouse. Based on a similar belief, Scallon sees the need for more university or in-house programs to train providers in understanding information drawn from EHRs and telehealth technology. Hyman thinks that independent practices would find value in a training program that would help them understand their practices’ data, workflows and transition to value-based health care.

Health care organizations stand at different points in their transformation to a data-driven care delivery model and culture. While they’ve encountered challenges in accessing and understanding data, they’ve also found ways to use BI or AI to improve operations and outcomes. Participants in the AHA’s data analytics panel discussion suggest that now is the perfect time to start having a conversation about analytics strategy — or better yet, begin documenting a formal strategy and plan.