My son Daniel was diagnosed with autism nearly 15 years ago when he was two years old. Back then, people thought that the rate of autism was 1 in 10,000 children, and the treatment options were limited. At that time, the first best hope for children with autism was to use some form of applied behavior analysis (ABA) based on the research done by Dr. Lovaas out of UCLA (Lovaas 1989). At that time, there were no ABA specialists in most major cities and few special schools for children with autism so we had to fly specialists in from California and Oregon to train local therapists and college students to work with our son.
One of the things we were taught by these ABA specialists was the importance of collecting data to allow us to better understand Daniel's response to therapy. This allowed us to understand how to teach him so that we could maximize his progress.
As is true for many children with autism, Daniel's learning patterns were not similar to any of the example cases we found or to any other children with autism that we knew. His ability to acquire spoken and even sign language was virtually nonexistent, and his only means of communication was to use pictures via the picture exchange communication system (PECS) (Frost & Bondy 1994).
Over the years we have collected data on all of his therapies, behaviors, and social skills. This data has always been valuable when it came to deciding what he needed to work on next in his therapy sessions. But even more importantly, it has provided a yardstick to measure the effectiveness of all of the interventions we have tried with him - and we have tried a lot of things!
When Secretin was touted as the next miracle cure for autism in the late 1990s, we were among the first to try it (e.g. Hovarth et. al. 1998). After the first round of shots, we were so hopeful. We were sure we saw improvements. But over the next few weeks our son's daily data showed us we were wrong. There was absolutely no change in his behaviors or in his rate of learning his academic objectives.
Later Daniel's behavior problems got out of control. He had frequent outbursts and would hit, kick, head butt and bite staff, students and us. It was not uncommon for him to commit 30 to 60 act of violence against others in a single day. The school had the district behavior specialist help them. We expanded Daniel's behavior data collection to include each behavior and trigger for his behaviors. We developed and implemented an individualized behavior plan in conjunction with the school. However, the data showed that, although interventions taken reduced injury to others, they did not reduce the number and intensity of the behavior episodes. The school finally suggested that we try some form of medication to reduce Daniel's violent behavior. When the first medication was tried everyone perceived an immediate benefit and felt that they were finally making progress towards controlling Daniel's behavior problems. However, after two weeks the behavior data collected at home and at school did not support this false perception. The initial medication was dropped and a second medication was tried. The benefit was immediate and dramatic. The number of behaviors dropped ten times within 3 days and continued to drop over the next few months. In this case, our systematic data collection and analysis allowed us to spot which behavior and medical interventions were working and to change course more quickly than if we were relying on gut instinct alone.
One of the challenges we faced when collecting and analyzing all of this data is the amount of time involved in taking the data on paper and then analyzing the data to help improve our decisions about Daniel's treatment. We would fill a three-inch binder (like the one in the picture below) with data every four to six months.
There are lots of things we could not do with this paper-based data. Generally, we used this data by scanning it to make quick calculations or inferences based on our general perceptions of the data for the last few days. Occasionally we would transfer some of that data to spreadsheets to create graphs to give us a "bigger picture" of Daniel's progress. But this took time so we never transferred all of the data into the spreadsheet program. There were other problems with the paper based data as well. The data was difficult to share. Typically, therapists only had access to it at our house and his teachers at school never saw it. It was also difficult to process the paper data to get a good picture of Daniel's long term progress. Finally, it was virtually impossible to find a comment or problem from a few weeks to a few months past to clarify what the problem was or how it was solved.
As a Professor of Information Systems at the University of Colorado Denver, I know the benefits of computer-based data collection and analysis systems have provided to business over the past 2 decades. When Daniel was nine, I finally decided to create a data tracking program for him. I wrote the first version of Daniel's Data Tracing program (DDtrac) in 2002. Over the past seven years, the program has gone through hundreds of changes as I thought of new things that I wanted to look at and new questions I wanted to be able to answer with his data.
Two years ago, with the help of some of my students from the University of Colorado Denver, I turned DDtrac into a web-based service that can be used for tracking data for children with a wide variety of special needs. The current system allows data to be collected for educational tasks, social interactions and behavior episodes. It automates many of the data analysis tasks that I used to spend hours performing by hand. And, it is on the web so we can easily share Daniel's data with his teachers and other specialists that work with him.
The biggest benefit we have found to having all of Daniel's data online is the ability to understand long term changes that are difficult to see without data. When Daniel was 12 we started him on a gluten-free/casein-free diet (Knivsberg et. al. 2003). Over a period of months the data we collected showed a definite improvement in his ability to learn new things. Heartened by the improvement and based on new research, we decided to institute an even stricter diet, the specific carbohydrate diet (Gottschall 2004). Data over the next few months showed an even greater improvement in Daniel's rate of learning and a dramatic reduction in his behavior problems (see picture below) and after six months on the diet, he began to understand and use words for the first time in his life. The changes we saw were so subtle at first, without the ability to chart his long term progress data we might have missed them. Since implementing a special diet requires a tremendous amount of effort for us and sacrifices for Daniel, it is likely that, without the evidence of progress, we might have abandoned the gluten-free/ casein-free diet after a few months and never have tried the more restrictive specific carbohydrate diet.
Over the years the therapies Daniel has received have changed as has the data we have collected. Using DDtrac has expanded our ability to use data to improve how people work with my son. It also allows us to generate charts to provide evidence to doctors and insurance companies about the benefits from different therapies and medications. Both of these things help us help Daniel. Today, as we start to add foods back into Daniel's diet so he can begin to eat more like a typical teenager, you can know that we will be watching his data closely to see if his rate of progress slows or if his behavior problems increase. For us, data has proven, on countless occasions, to be one of the best tools we have had for improving outcomes for our son.
For more information on the DDtrac software discussed in this article, please go to: developingmindssoftware.com.
________________
Dr. Dawn G. Gregg is an Associate Professor of Information Systems at the University of Colorado Denver and cofounder of Developing Minds Software. Her research focuses on how to organize and maintain Web-based content so that it can be used to better meet special education needs. Her work has been published in numerous journals such as MIS Quarterly, Communications of the ACM, and Decision Support Systems.
____________
References
Frost, L. A. and Bondy, A. S.
The picture exchange communication system: Training manual. Newark, DE: Pyramid Educational Consultants, 1994.
Gottschall, E.
Breaking the Vicious Cycle: Intestinal Health Through Diet, Baltimore MD: The Kirkton Press, 2004.
Horvath K, Stefanatos G, Sokolski KN, Wachtel R, Nabors L, Tildon JT, "Improved social and language skills after secretin administration in patients with autistic spectrum disorders,"
Journal of the Association of Academic Minority
Physicians, 9 (1), 1998, pp.9-15.
Knivsberg, A.M., Reichelt, K.L., Høien, T., and Nødland, M. "Effect of dietary intervention on autistic behavior,"
Focus on Autism and Other Developmental Disablities,
18 (4), 2003, pp. 247-56.