Kessel and Combs (2016) published a compelling article about the developments in electronic data management within healthcare. We’d like to summarize and apply their findings for today’s Applied Behavior Analysis (ABA) community, which is increasingly reliant on the use of electronic data collection. The study investigated the results of 895 articles published over the last decade to examine the massive growth in healthcare data and the necessity of harnessing and sharing “Big Data” (Kessel & Combs, 2016).
Mining big and small data
While various healthcare providers define “Big Data” differently, a general guideline for big data is when you have more data points than fit on the one million rows of an excel spreadsheet. Thus far, ABA has been able to mine “small data” for an incredible abundance of behavioral insights with real world predictive capacity. This is of particular import, as single-subject designs have previously been more commonly utilized in evidence-based practice (Horner et al., 2005). While group experimental designs allow for normative comparisons, single-subject designs enable the reader to analyze data at an individual level (Morgan & Morgan, 2009). Socially significant applied behavior change is a cornerstone of behavior analysis (Baer, Wolf, & Risley, 1968).
Tapping the Quantified-self movement
Big data offers ABA a new opportunity to produce incredible results, especially in the area of special needs. Every child with special needs is different, but a large aggregated ABA database can combine the results of intervention for millions of children, especially with regards to path-of-treatment and personalized care. Furthermore, the quantified-self movement (e.g. movement trackers and sleep trackers for your phone) offers us a chance to correlate general health results with ABA results. For example, wouldn’t it be amazing to know how many hours of sleep your kiddo got last night? Or how specific medications affect behavioral intervention results?
Building isolated vs collaborative data systems
Kessel and Combs (2016) stress the need to be vigilant in how we manage this revolution. Collaboration between medical specialties is key to interdisciplinary research and personalized medicine, but to do so we must have standardized and shared data. In healthcare, clinical facilities typically develop two systems for managing information. One handles daily treatment and patient documentation, while the other handles scientific content including clinical trials and research. These are usually (but not always) built to leverage data and work together.
Unfortunately, the two systems and the connection between them are often developed by researchers in-house to meet the isolated needs of the organization, without regard to sharing, interoperability, or future growth. It is imperative that Big Data systems be implemented using standard communication protocols because the future requires a much larger collaboration effort than any organization’s walls can hold. Further complicating matters is the fact that we need many data tools to connect and provide a minimum quality of data for the big data approach to be worthwhile.
The quest for standardized data collection
As of yet, the field of ABA has no electronic standard for data collection such as discrete trial instruction (Smith, 2001), an instructional evidence-based methodology that is frequently implemented with learners diagnosed with autism. This lack of standardization in electronic data collection presents its own impediments within our industry. Thread Learning® provides a Microsoft® Excel data dump free to all customers, but we are still working to figure out the best format to give our customers. We hope to engage other software companies in standardizing ABA data and sharing it freely between systems.
Are we ready for the future?
Is healthcare ready for big data? Kessel and Combs (2016) say perhaps not yet, but the technology is available and the future is bright. From Electronic Data Collection to Practice Management and ABA Training videos, we think applied behavioral analysis is also marching slowly but surely towards a bright data future.
Correspondence
Greg Brill: greg@threadlearning.com
Noor Syed: nos218@lehigh.edu
The authors acknowledge Sue Lacy for editing and review.
Review of Developments in Electronic, Clinical Data Collection, and Documentation Systems over the Last Decade — Are We Ready for Big Data in Routine Health Care?
_By Kerstin A. Kessel, Department of Radiation Oncology, Technische Universität München, Munich, Germany. Stephanie E. Combs, Institute of Innovative Radiotherapy (iRT), Helmholtz Zentrum München, Neuherberg, Germany.
Published: Front. Oncol., 30 March 2016_ https://doi.org/10.3389/fonc.2016.00075
Big Data photo by Joshua Sortino _on_ _Unsplash _
Greg Brill, CEO, Thread Learning and Noor Syed, Ph.D., BCBA-D, Lehigh University Autism Services Ph.D. (July 7, 2018)