A 5-Year Update of Conflicts of Interest in Autism Intervention Research in Applied Behavior Analysis Journals

Abstract

Researchers’ false, incomplete, or missing disclosures of conflicts of interest (COIs) can introduce bias into research, can erode public trust in research findings, and represent ethical violations of most academic journal policies. A 2020 study discovered that publications in applied behavior analysis (ABA) journals are particularly problematic in adherence to COI disclosure ethics. The current study is a 5-year update of this previously conducted study. We examined autism intervention research articles published over a 1-year period in eight ABA journals. Two coders extracted author names and COI disclosure statements from each study and conducted web searches to determine if authors were affiliated with organizations providing ABA services or consulting. One hundred and nineteen studies met our inclusion criteria, from which we compiled a database of 450 authors. Seventy-eight percent of authors held clinical and/or consultancy COIs. At the study level, 93% of studies were written by at least one author with a clinical and/or consultancy COI. Only 8% of studies disclosed any author COIs, and only 2% disclosed clinical and/or consultancy COIs. Ninety-three percent of statements claiming no COIs were false. COIs are increasingly pervasive in ABA autism intervention research, and the vast majority remain undisclosed.

Lay Abstract

This study looked at how often researchers who publish about autism interventions in journals focused on one type of intervention called Applied Behavior Analysis (ABA) tell readers about their conflicts of interest (COIs). COIs happen when researchers benefit from showing something specific in their research, such as an intervention making things better for autistic people. The COIs we looked at are when researchers also receive money to provide ABA to autistic people or help other researchers provide ABA to autistic people (i.e., they worked as a consultant). COIs can negatively affect how research is designed, interpreted, and presented. We wanted to see if researchers tell readers about their COIs, or if they say they do not have COIs when they do. We reviewed autism-related intervention papers published over 1 year in eight ABA journals. For every paper, we copied the COI statement. Then, we searched online to see if authors were working as or consulting with ABA service providers. We looked at 119 papers with a total of 450 authors. This study is a five-year update of a 2020 study that found widespread but rarely reported financial COIs among ABA researchers. In our updated study, we found that 78% of authors had a COI. Some worked in ABA clinics, some offered paid consulting to other ABA providers, and some did both. Almost all papers (93%) had at least one author with these kinds of connections. But very few (8%) mentioned any COIs, and only 2% of papers stated that the authors worked as ABA providers or consultants. Most papers said the authors had no conflicts at all, but this was often not true. In fact, 93% of “no COI” statements were false. Although more ABA journals now require disclosure than in the past, many statements are still inaccurate, showing that the problem has not improved. The people in charge of publishing research, and the people who write research papers, need to do much better to let readers know about researchers’ COIs.