Monday, July 20, 2009

Whitworth professors, students develop software to support autism identification, treatment

Vision includes making system available online for parents, physicians

The number of children diagnosed with Autism Spectrum Disorder (ASD) this year will exceed those diagnosed with AIDS, diabetes and cancer combined; experts say one out of every 150 8-year-olds is affected by the disorder. Due to the complex nature of ASD, however, many pediatricians don't have sufficient training to confidently screen patients for it. What's more, once a child is diagnosed, parents and professionals must choose from hundreds of available treatments, some helpful, but many worthless.

To advance the study and treatment of this growing epidemic, Susan Mabry, Whitworth associate professor of mathematics & computer science, and Betty Fry Williams, Lindaman Chair and professor and coordinator of special education at Whitworth University, have teamed up to create a highly-specialized analytical software application that will screen for ASD, help doctors make referral decisions, and track patients' progress.

"It is our contention that automated software promises viable methods for analyzing varied dimensions, treatments and causes of ASD," Mabry says. "We are confident that a close affiliation between autism experts and computer scientists is the only avenue that will produce answers for this devastating disease. Through this research, we hope to make meaningful contributions to the growing body of autism findings and advanced computer science methodologies."

Mabry and Williams envision that the system will be accessible online so parents and professionals can add their own assessment data and autism patients' progress can be documented over time.

"We're hoping this will help us verify the treatments that create the best outcomes for children with ASD and will perhaps further identify which symptoms respond best to certain treatments," Williams says.

ASD is a serious and complex neurological disorder that impedes a person's ability to communicate and relate socially to others, Williams says. Symptoms include impaired communication, reduced social interaction, and preoccupation with limited items or topics. More than 500 treatment options exist, ranging from special diets to swimming with dolphins, which create difficult decisions for parents of children with autism, she says.

Mabry and Williams decided to conduct the research after Mabry, who has a background in medical informatics, attended the Lindaman Chair lecture Williams gave at Whitworth last year about research on the causes of ASD. Mabry's previous research project, which was funded by the National Science Foundation, was an intelligent agent clinical decision support system that targeted critical care and trauma environments. After hearing Williams talk about the complexities involved in autism detection and treatment, Mabry wondered whether she could adapt and expand her previous project to address the problems associated with screening and treating autism.

Williams says when she heard Mabry's idea, she immediately thought of a recent recommendation by the American Association of Pediatrics that pediatricians should screen all children 3-years-old and younger for symptoms of autism and should refer them for treatment when appropriate. Yet many pediatricians don't have the specialized training or experience necessary to carry out that screening process, Williams says.

"The variance within the spectrum, breadth of severity levels, extensive types of data, and numerous potential causes, as well as overlapping factors and diagnosis, all contribute to an immensely complex problem," Mabry says. "It is all but impossible to manually sift through expansive data to reach meaningful conclusions, which is why we believe automated software holds the most promise."

The project involves three distinctive software systems, which are in various stages of development. The first project Mabry and her computer science students began working on last summer, supported by summer research student fellowships from Whitworth, involved developing a screening tool to identify the possibility of autism in children and to suggest further evaluation if warranted. Using the computer protocol, pediatricians can gather information about symptoms that are related to ASD. Weights are assigned to each symptom and the software analyzes the data to determine if a child appears to be developing normally or should be evaluated further for ASD. Mabry and the students expect to wrap up the project this summer, and they hope to apply for grants to have the weighting system further refined by specialists and then field tested by pediatricians.

The second project involves developing a tracking and analytical tool that would trace patients' treatments and progress to identify correlations between the two, thereby identifying which treatments might be most effective. A team of computer science students led by Alice Clawson, '09, began the project this past spring. Clawson presented the team’s work at the Northwestern Association for Behavioral Analysis Regional Conference, held at Gonzaga University in March. Mabry and one of her students are developing it further this summer through a Whitworth student research fellowship.

The third, planned project is an ambitious ASD agent mining system that will adapt Mabry's previous intelligent medical agent system to analyze large data banks, with the goal of identifying common causal factors and trends of ASD. Mabry and Williams have recently joined the Interactive Autism Network Research Community and have received approval to tap into their large data bank, hosted by the Kennedy Krieger Institute at Johns Hopkins University, and are positioning to pursue national and private funding for this project.

Williams, as well as Dana Stevens, an instructor in education at Whitworth, are serving as consultants to Mabry and her students, and they're connecting other ASD experts in the Spokane community to the project as well.

Located in Spokane, Wash., Whitworth is a private liberal arts university affiliated with the Presbyterian Church (USA). The university, which has an enrollment of 2,600 students, offers 53 undergraduate and graduate degree programs.

Contacts:

Susan Mabry, associate professor of mathematics & computer science, Whitworth University (509) 777-4686 or smabry@whitworth.edu.

Betty Fry Williams, Edward B. Lindaman Chair and professor of education, Whitworth University, (509) 777-4688 or bwilliams@whitworth.edu.

Emily Proffitt, public information officer, Whitworth University, (509) 777-4703 or eproffitt@whitworth.edu.