Addressing Common Problems in Autism Research: Cohort Development and Sample Description for a Participatory, Real-World, Longitudinal Study of Outcomes in Autistic Adults.

Adult autism services research is plagued by methodological challenges, including difficulty obtaining representative or heterogeneous samples and a frequent mismatch between experimental methodologies and real-world settings. The Academic Autism Spectrum Partnership in Research and Education (AASPIRE) Outcomes Project used a community-based participatory research (CBPR) approach to (a) create the AASPIRE Measurement Toolkit - a set of self- and caregiver-reported accessible outcome measures to evaluate health and social services and (b) explore factors predicting changes in outcomes. We previously used a CBPR-nested Delphi process to identify high-priority outcomes and adapted or created instruments to measure them. We conducted a longitudinal survey to validate these measures in a pragmatic sample of autistic adults with subcohorts from two health care systems, two disability service systems, and the larger autistic community in the United States. This article (a) describes our study methods and the sample's baseline characteristics and (b) compares characteristics among participants in the separate subcohorts and sites. We collected baseline data from 870 participants. Follow-up response rates were >80% for the second and third time points. The sample includes strong heterogeneity. As expected, sample characteristics differed markedly between the three subcohorts. Lessons from our experience may help other researchers devise strategies to effectively recruit heterogeneous samples of autistic adults across various settings.Lay AbstractThere are many challenges that can get in the way of doing good research on autism in adulthood. Examples of these types of challenges include the following: (a) it can be hard to include autistic people with a wide range of characteristics in research studies and (b) traditional research may come up with results that do not really apply to people in the real world. Academic Autism Spectrum Partnership in Research and Education (AASPIRE) is a group of academics and autistic community members who work together as equal partners. We are creating the AASPIRE Measurement Toolkit, an accessible set of survey measures about the health and life outcomes most important to autistic adults. In an earlier part of the project, we worked together to create the measures in the Toolkit. Now we are doing a big survey study to test the new measures and look at what predicts changes in outcomes over time. This article describes our study methods and the characteristics of the people in the sample. It also compares basic information about people recruited to the study from different places, using different methods. We recruited 870 autistic adults from two health care centers (the health care subgroup); two disability services systems (the disability subgroup); and the broader autistic community. We collected data from each person three times over 12 to 18 months. Over 80% of study participants did the follow-up surveys. People in the study had a wide range of personal and disability-related characteristics. As we predicted, the three subgroups were really different from each other. Even the two sites within the subgroups were a bit different. This serves as a warning to researchers: results that come from people recruited from one setting may not apply to people in other settings. It also shows how using multiple different strategies at the same time can increase the diversity in study samples.