Electronic Health Records

Integration & Interoperability News

States Could Utilize AI for Improved Healthcare Interoperability

Different states are at varied levels of data integration and healthcare interoperability, but advanced data analytics capabilities, including AI, could fuel improvement.

healthcare interoperability artificial intelligence

Source: Thinkstock

By Elizabeth Snell

- There are numerous obstacles to seamless healthcare interoperability at state health and human services (HHS) enterprises, but utilizing new data analytics options could prove fruitful in improving patient outcomes, recent research found.

Specifically, artificial intelligence (AI) solutions may provide HHS enterprises with greater insights into how to improve population health, decrease costs, as well as reduce substance use disorders, unemployment, and homelessness.

Researchers at Leavitt Partners interviewed HHS officials in Colorado, Idaho, Oklahoma, Utah, and Washington. The states were at different levels of “leveraging data and pursuing interoperability to inform state policies and improve outcomes,” the team explained. 

“Data, and the ability to share data, sits at the center of these cross-governmental and health care collaborations,” researchers explained. “Data has the potential to connect distinct hospitals, clinics, government agencies, and community-based organizations and provide a more complete and accurate picture of an individual and their needs.”

One of the top challenges for HHS enterprises is that they tend to have complex organizational structures, and they utilize programs operating under multiple state agencies (or divisions within umbrella agencies), local social service organizations, and health care settings, researchers explained.

READ MORE: Recognizing Healthcare Interoperability as a Moving Target

Therefore, structures have “evolved under separate systems operate under different regulatory and governance structures, and utilize varied IT infrastructure.”

“This creates unique challenges involving the data itself, and introduces a variety of secondary considerations including governance, privacy, regulations, and financing,” wrote researchers. “Unlocking data’s potential means ­first understanding and then addressing these challenges.”

Various state organizations also do not collect the same amount or type of data, and processes must be implemented to ensure data integrity, standardization, and accuracy. HHS enterprises may also define terms differently or use different claim identifiers, the researchers noted.

Interoperability is a key way to overcome such data fragmentation, but the variety of IT systems, diverse levels of sophistication, and varied interoperable capabilities also make seamless connectivity difficult.

“Challenges also exist with data governance, interagency relationships, and developing common data de­finitions across the state HHS enterprise,” the team wrote. “Legal, regulatory, ethical, and political concerns around data ownership, security, and use can impede an agency’s willingness or ability to share their data across the state HHS enterprise.”

READ MORE: Allscripts Debuts Machine Learning, Cloud-Based EHR System Avenel

Financial challenges can also impede improved data system capabilities. For example, enterprises may need to upgrade or replace old IT systems, train staff, or hire new staff who are already knowledgeable on new analytic capabilities.

“Even when a state has money, it can be difficult to make the case for large investments in the IT department over more programmatic spending,” researchers explained. “Overcoming resource barriers requires a tangible vision for how investment in analytics and interoperability can improve outcomes.”

Even so, researchers found that the five reviewed states are making efforts to improve healthcare interoperability, showing that different strategies could positively impact the population.

Washington is using predictive analytics to improve outcomes, the team found. The Washington State Department of Social and Health Services (DSHS) has a longitudinal client database that pulls and uses data from over 30 systems both inside and outside of DSHS.

Predictive analytics also allows the state to determine that foster youth who are parents, or who have been homeless, or who are African American have the greatest risk for experiencing homelessness after aging out of foster care,” researchers said. “Based on these predictive analytic capabilities, the state can develop more targeted, more effective interventions.”

READ MORE: New York PDMP Achieves Healthcare Interoperability with 25 States

Additionally, the Colorado Department of Healthcare Policy and Financing (HCPF) is using champions to drive data analytics and interoperability efforts.

“Colorado developed an integrated eligibility system that allows for real-time eligibility assessment across 20 programs, such as Medicaid and Temporary Assistance for Needy Families (TANF),” the research team stated. “HCPF [also] has the data and the analytic capabilities to run predictive analytics to assess the likely trajectory of an individual’s health condition.”

HHS enterprises are at different stages of utilizing data analytics options, but Leavitt maintained that there is a consistent push among all of them to implement more advanced technology. There must be continued development of human and technological capabilities though, researchers stressed.

“The trend toward next generation technology on the data analytics spectrum appears to be artificial intelligence,” the team wrote.

“Unlike traditional data analytic systems, cognitive computing can consume both structured and unstructured data and present it in ways that are actionable for decision-makers,” researchers continued. “Moreover, similar to the human brain, cognitive computing systems possess the capacity to improve and develop over time with the input of additional information.”

Implementing AI will likely bring forth its own set of challenges, the report acknowledged. States that are in early stages of data interoperability may not have the required IT system capabilities and enterprise-wide data management strategy for a cognitive solution.

Additionally, staff members might not have the necessary training or knowledge of AI to properly utilize it.

Data accuracy will also be a challenge. A 2017 US Department of Health and Human Services report explained that AI application development “will perform poorly when signi­ficant data streams are absent.”

Overall, it is worthwhile for HHS enterprises to pursue cognitive computing and other AI systems, the team maintained.

“HHS of­ficials have access to—and are currently pursuing—a variety of strategies to overcome these challenges,” researchers concluded. “Although states are in various stages of data integration and interoperability, potential exists for data systems to provide the most holistic view possible of an individual, predict an individual’s risks, and possibly inform measures to intervene before a risk is realized.”



Sign up to continue reading and gain Free Access to all our resources.

Sign up for our free newsletter and join 60,000 of your peers to stay up to date with tips and advice on:

EHR Optimization
EHR Interoperability

White Papers, Webcasts, Featured Articles and Exclusive Interviews

Our privacy policy

no, thanks

Continue to site...