Frank Cicero, PhD, BCBA, LBA
Seton Hall University

Having been asked to review published articles for ASAT’s library of Research Synopses in the past, I always enjoyed having the chance to learn more about what is going on in autism research, particularly outside of my area of specialization. This was the case when I was asked to have one of my graduate students at Seton Hall University review the article Randomized controlled trial of vitamin D supplementation in children with autism spectrum disorder by Saad et al. (2018). According to the authors, the rationale for the investigation into the role of vitamin D in the treatment of autism spectrum disorder (ASD) was related to the role vitamin D plays in brain homeostasis, neurodevelopment, aging, and gene regulation. More specifically, the study was undertaken due to a reported link between maternal vitamin D deficiency during pregnancy and an increased risk of ASD diagnosis. According to previous data reported by Saad et al., 57% of a sample of 122 children with ASD had a vitamin D deficiency and an additional 30% had vitamin D insufficiency. They therefore decided to conduct a randomized clinical trial to evaluate the effects of vitamin D3 supplements on the core features of ASD.

After completion of the research, the study was ultimately published in The Journal of Child Psychology and Psychiatry in January of 2018. In reviewing the methods section, the study seemed like a well-designed and well-conducted experimental investigation. The study was a double-blind, randomized, placebo-controlled clinical trial that was conducted over a 4-month period. Participants included 120 children with ASD receiving services in outpatient clinics and private autism treatment centers in Egypt. ASD diagnoses were confirmed through DSM-5 criteria, a detailed history interview, and a 3-hour observation using the Childhood Autism Rating Scale (CARS). Participants were randomly assigned to two groups, one receiving vitamin D3 supplemental treatment through liquid drops (experimental group) and one receiving matching placebo liquid drops (control group). After four months, the children were assessed using scores on four autism questionnaires as outcome measures: the Aberrant Behavior Checklist, the Childhood Autism Rating Scale, the Autism Treatment Evaluation Checklist, and the Social Responsiveness Scale.

This sounded like the makings of a very impressive study. It was interesting and exciting to find that the results indicated a statistically significant improvement on all four outcome measures in participants in the experimental group as compared to the control group. The authors’ hypothesis was supported. The study was the first published, double-blind, randomized clinical trial showing the efficacy of vitamin D supplements in children with ASD.

Prior to submitting our article synopsis to ASAT for publication in Science in Autism Treatment, I always do a final review to make sure no modifications are required. Luckily, I did not make an exception in this case. To quickly re-read the original article, I attempted to pull up the article through Google Scholar. To my surprise, I did not find the Saad et al. (2018) article that we had just reviewed. Instead, I found an entry in the Journal of Child Psychology and Psychiatry, published in June of 2019, stating that the Saad et al. (2018) article had been retracted by the journal’s Editor-in-Chief. The statement explained that the journal had received several letters to the Editor highlighting potentially questionable statistical procedures used by the authors. These concerns, if valid, would have brought the overall findings into question. This prompted the journal Editor to request the raw data used in the analysis. The authors provided a reply that caused more concern; they were only able to submit partial data for review because of a computer outage that resulted in the loss of their research database. A re-analysis of the partial data revealed a significant discrepancy to what was reported by the authors. The journal also uncovered previously unreported missing data and recording issues associated with changes in the treatment regimen over the course of the study. As a result of these concerns, the journal decided that a retraction of the original Saad et al. (2018) study was necessary.

Once again, it was lucky that I caught the retraction prior to submitting the article synopsis to ASAT. In this case, the peer-review journal publication process appears to have worked without resulting in harm, or did it? Keep in mind that the original article was published in January of 2018. It was not until June of 2019, 17 months later, that the article was officially retracted. As with my student and I, it is likely that many others also read the original Saad et al. (2018) article, walking away with the conclusion that significant improvements in autism symptoms were the result of four months of treatment with vitamin D3 supplements; a conclusion that is now in question and should not have been published. Unlike my student and I however, it is unlikely that most other readers were informed of the subsequent retraction. Although the journal can retract studies of concern, they cannot erase the belief in false conclusions that is left in the minds of readers who are not aware of the retraction.

Herein lies the potential danger of not being aware of the peer review process and the concept of retraction. This occurrence, therefore, has prompted us to write up a brief review of the process of study retraction and the role it plays in the shaping of evidence-based practice, which will appear in the next issue of Science in Autism Treatment.

Citation for this article:

Cicero, F. (2020). Perspectives: An unexpected journey into retraction. Science in Autism Treatment, 17(9).

Related ASAT Articles/Resources:

Print Friendly, PDF & Email