The National Database for Autism Research (NDAR) is a collaborative biomedical informatics system sponsored by the National Institutes of Health. Primary objectives of NDAR are to provide qualified researchers with access to high quality, detailed human subjects data that underlie research findings relevant to autism spectrum disorders (ASD) and to facilitate data sharing across research projects. Sharing of research data provides the opportunity for researchers to (a) use defined standards to characterize their data; (b) validate results; (c) more easily pool relevant research data acquired by others; and (d) communicate results at the appropriate time with other researchers and, if appropriate, the general public.
Objectives:
The objective of this analysis is to (a) characterize the type of research being conducted by the ASD research community to understand how NDAR may contribute to accelerating ASD research discovery and (b) provide an objective measure on the research community's progress in increasing the study sizes relevant to ASD research.
Methods:
2008 and 2009 International Meeting for Autism Research (IMFAR) presentation abstracts (779 in 2008 and 842 in 2009) were reviewed. Data collected include:
- Whether the study included human subjects or biomaterials
- Sample size of enrolled human subjects with ASD as well as controls
- Whether the study focused on NDAR relevant categories of treatment/intervention, phenotyping, genetics or neuroimaging/EEG
A priori exclusion criteria were established to confine subsequent analyses to studies most appropriate for NDAR. The exclusion criteria were that the research abstract:
- Did not involve human subjects or the number of human subjects was not reported
- Focused on the development or validation of measures/assessments
- Involved cell/animal models
- Involved a literature review, bioinformatics system review, or survey of treatment providers
- Reported on a self-selected survey population
- Had an epidemiology focus
Results:
510 (65%) abstracts from IMFAR 2008 and 574 (69%) abstracts from IMFAR 2009 were included in this analysis. In total, the average proband sample size increased from 67 to 106 subjects from 2008 to 2009. Likewise, median sample size increased from 24 to 30 subjects during this period. Standard deviation almost doubled from 118 to 330.
| Category | Num '08 | Num '09 | Mean '08 | Mean '09 | SD '08 | SD '09 |
| Treatment/Int. | 70 | 65 | 31 | 40 | 71 | 52 |
| Phenotyping | 336 | 382 | 60 | 100 | 99 | 340 |
| Genetics | 47 | 38 | 218 | 405 | 219 | 578 |
| Neuroimaging/EEG | 57 | 89 | 21 | 29 | 15 | 29 |
Conclusions:
This comparative analysis provides evidence that sample sizes for ASD studies are growing. However, relatively small samples sizes are the norm in ASD research and a large portion of ASD studies focus on phenotype data. Of particular note are the small samples sizes for treatment/intervention studies.
We suggest that the increasing ASD research community adoption of large, well characterized, high quality datasets (e.g. Simons Simplex, CPEA/STAART, AGRE) may be postively impacting the growth in average sample sizes and skewing the standard deviation in the phenotyping, genetics, and neuroimaging/EEG categories. In future years, we will evaluate the long-term impact of linking/including these datasets with NDAR in an effort to increase substantially the study sample sizes of research presented at IMFAR. It is our plan to continue to monitor this trend.