Machine learning detects language disparities to improve COVID-19 tracing


Among the challenges posed by the pandemic were the language barriers public health agencies faced as they struggled to trace the spread of infection among Latin American communities.

In Santa Clara County, California, only 25% of the population is Latino, but that accounted for over 56% of the state’s COVID cases. This has put Spanish-speaking contact tracers – who call patients with diagnoses, identify and notify their contacts, and help with isolation and quarantine – in high demand.

These Spanish-speaking contact tracers have been essential in reaching potentially infected people as quickly as possible, but with thousands of cases per day – and a limited number of Spanish speakers and interpreters – it can take days to alert contacts in the world. ‘a patient.

An additional challenge is that Spanish-speaking residents may be reluctant to speak with government employees asking them for complex personal information, especially through someone who is not fluent in the language, according to a report in Human-Centered. AI News from Stanford University.

To improve contact tracing in the Latino community, Santa Clara County health officials have partnered with experts from Stanford University’s RegLab, a group that designs and evaluates programs, policies and technologies. to modernize government. They wanted to see if they could predict when a contact only speaks Spanish or has limited fluency in English. With this idea, the county could then assign the patient to one of the county’s native Spanish speakers.

A team led by Daniel Ho, faculty director at RegLab and associate director of the Stanford Institute for Human-Centered Artificial Intelligence, used machine learning to predict people’s language needs, helping contact tracers resolve larger cases. quickly and reduce the health gap between Latinos in the county and other communities.

Contact tracers usually start with the most basic information about the people they call, such as the patient’s name, address, date of birth, and test result.

The researchers combined this rudimentary data with demographic information from the census and other administrative data. A machine learning algorithm analyzed and weighted data such as census block group, age, and race and ethnicity information from census and mortgage data and identified patterns that predict language preference. Contacts were scored as to what language they would likely prefer before being assigned to a plotter.

To test the effectiveness of the algorithm, RegLab worked with Santa Clara County to perform a test that randomly referred half of the cases to a “language specialty team” with bilingual speakers and processed the other half with the typical county process.

In just two months, the benefits became evident. In the test group, the time it took to complete cases decreased by almost 14 hours compared to the control. Same-day completions rose 12% and the number of people refusing to be interviewed fell 4%.

“Based on the results and success of this trial, Santa Clara County has extended language matching to all [Santa Clara Public Health Department’s Case Investigation and Contact Tracing], and the state of California is considering adoption into the statewide system, ”the authors said in their article.

“When we connect with people in their preferred language, it makes a huge difference in their willingness to share information about themselves, their health, and their families and friends,” said study co-author. , Dr. Sarah Rudman, Director of Contact Tracing for the Santa Clara County Public Health Department.

The new approach not only improved people’s willingness to engage in the process, but it also allowed Rudman to ensure that bilingual county plotters could be assigned to the contacts most likely to need them.

“Before the algorithm, you could hear the frustration in the voices of our plotters when they didn’t match a contact,” Rudman said. “After the algorithm, we would talk about the families they had been connected with, many of whom stayed on the phone only because the tracker spoke Spanish and pronounced their names correctly.”

When each missed contact can mean additional infections, those are significant improvements, Ho said, noting that the partnership between the people and the machine was a surprising – and refreshing – outcome for him.

“The AI ​​community is very concerned about whether machines are going to replace human judgment,” he said.

The full study is available in the Proceedings of the National Academies of Science.

Source link


About Author

Comments are closed.