Effective use of predictive analytics to improve health outcomes in Hawaii will require breaking down communication barriers


Shane Ersland


Establishing clearer communication standards will be key for health professionals in using predictive analytics to improve health outcomes for Hawaiians.


Stay one step ahead. Join our email list for the latest news.



Industry experts discussed efforts to utilize predictive analytics during the 2023 Hawaii State of Reform Health Policy Conference last month. 

Derek Vale, health systems management office chief at the Hawaii Department of Health’s (DOH) Behavioral Health Administration, said DOH is currently considering situations in which predictive analytics can be used to improve health outcomes, including preventing readmissions, maintaining population health, increasing patient engagement and outreach, and speeding up the claim submission process.

“One of the biggest issues we’ve seen on the state side is the completeness of the data and data standards,” Vale said. “One of the biggest challenges we have as a public health entity is understanding where the different data sources are. And whether everyone is looking at those data standards the same way.”

Providers can encounter barriers when attempting to use predictive analytics to determine standards for race and ethnicity, Vale said. 

“There’s a new initiative in DOH, and we’re hoping to work with other partners in the community in establishing more clear and representative standards for our community,” he said. “What it’s going to take for everyone to agree to those standards is anybody’s guess. What is it going to take for us all to adopt the same standards so we can use the data as value?”

Al Ogata, chief technical officer at the Hawaii Health Information Exchange (HHIE), said HHIE staff have raised questions about race and ethnicity information when looking at data from providers to try to determine existing health inequities.

“I think the number was in the 70-74% range where we didn’t see any information on race, gender, and ethnicity,” Ogata said. “When (staff) investigated, they found it wasn’t a technical issue. It was just that when the information was reported, it just wasn’t there. The information wasn’t made available to the facility, either because someone didn’t ask for it or the patient said [they] didn’t feel comfortable sharing.”

There are many times when health professionals are asked to gather information that can be sensitive to patients, Ogata said. 

“And they may or may not want to disclose it for certain reasons,” he said. “And that’s really the kind of thing we have to work through. Do they trust who they’re giving it to? It does bring in things you’ve heard about social determinants, and the fact that we need to get information on other things besides clinical information in order to help people be healthy.”

Dr. Monique Chyba, mathematics professor at the University of Hawaii, said the COVID-19 pandemic presented some challenges around using predictive analytics as well. She said the only source she had to utilize while studying the coronavirus with her students was the DOH’s data dashboard. She did not know where to get information about other important factors, like which hospitalized patients had been vaccinated.

“Predictive analytics need a very robust set of data so you can train the algorithm,” Chyba said. “For COVID, we took a different approach. We didn’t have access to data other than the dashboard. So we took a much more classical approach. It worked, but it was very tedious.”

Ogata said COVID taught HHIE that there are many barriers that need to be broken down in order for health professionals to effectively share data.

“How do we get more organizations to share information?” he asked. “How do we get them to share at the right level so that whether you’re a healthcare clinic or a payer, the information you’re exchanging makes sense?”