The National Institute of Mental Health created the National Database for Autism Research (NDAR) to accelerate autism knowledge through data sharing and collaboration. However, our experience using NDAR reveals systematic challenges across several aspects of data submission, selection, management, and analysis that limit utility of this resource. We describe our NDAR experience in an ongoing project examining autism intervention outcomes among marginalized racial, ethnic, and gender groups. For this study, we planned to gather data from NDAR to conduct an individual participant data meta-analysis. Eighteen studies met inclusion criteria and reported data on participants at more than one point in time on the Vineland Adaptive Behavior Scales (Vineland) and the Autism Diagnostic Observation Schedule (ADOS). The difficulties with submitting, selecting, downloading, and managing data from NDAR posed limitations on data availability and analysis. Of the 3,850 unique participants in the selected studies, data at multiple time points were available for 312 participants on the Vineland and 278 on the ADOS. No participants had data on all assessment domains. To accelerate autism research via data sharing and collaboration with NDAR necessitates improving the processes for submitting, selecting, and managing data.