Abstract
In the past 10 years or so, there has been a dramatic increase in research devoted to understanding outcomes for adults with autism. A confluence of factors has influenced this change. Parents who so fiercely advocated for their children 20 years ago are continuing to effectively advocate as their children are now entering young adulthood. Adding to the voices of parents are the increasingly influential voices of autistic self-advocates, who articulate the needs and desires of their community. Furthermore, it has become increasingly apparent that the adult service system is inadequate to meet the needs of adults with an autism diagnosis now, much less the increasing numbers who will be exiting high school with an autism diagnosis in the next decade. These concerns have been mirrored in the empirical literature: to date, nearly every study finds that adults with autism have difficulty accessing disability and medical services, experience high rates of unemployment and underemployment, face difficulties in daily living skills and achieving independence, and contend with elevated rates of physical and mental health problems (e.g. Croen et al., 2015; Hofvander et al., 2009; Howlin et al., 2004; Lugnegard et al., 2011; Shattuck et al., 2012, 2011; Taylor and Seltzer, 2011).
Yet, despite this compelling evidence for the challenges faced by many adults on the autism spectrum, we have learned disappointingly little about how to better support these individuals in reaching their maximum potential. One of the major roadblocks to making progress in this area is perhaps one of the most basic—there is little consensus on what outcomes are actually good outcomes and thus should be the target of services and intervention.
Given the many potential areas of need, it is unclear which potential aspects of adult life, if altered, would have the greatest impact on well-being and quality of life. If an adult with autism has few friends, does this indicate a “bad” outcome? Given the heterogeneity of autism, it is likely that some adults could be distressed by this situation whereas others would find it satisfying. Is paid employment always an indicator of a “good” outcome? Or is the fit of the vocational activity to the interest and skills of the individual more important? Should community living be considered a “good” outcome for all? As researchers, we often make value-based decisions about what activities we consider to be a good outcome, yet the very same activity or type of activity might have a different meaning to adults across the autism spectrum.
Perhaps then, one should take a personalized approach, allowing each person and family to define which outcome is the most important to target for that person. However, that approach is problematic for research purposes, when some degree of commonality and standardization is necessary. It may be that a hybrid of standardized and personalized approaches will be most helpful when defining “good” outcomes—particularly given the vast heterogeneity observed across numerous domains among adults with autism. One potentially helpful example of a hybrid approach is Goal Attainment Scaling (Cytrynbaum et al., 1979; Oren and Ogletree, 2000; Ruble et al., 2012); when used in an educational setting, Individualized Education Plan goals are collected and then progress toward each of those goals is rated on a standardized scale. A similar approach could be used in adult outcome research, whereas an individual and his or her family and other important supporters could develop individualized goals, and then a “good” outcome would be determined by measuring in a standardized way whether those goals have been met or exceeded.
Another way to personalize our definition of a “good” outcome could be to focus less on objective indicators such as employment, independent living, or relationships, and more on subjective indicators like quality of life or happiness. Subjective quality of life, in particular, is a promising area of research, and a growing number of studies are examining the factors associated with better quality of life for those with autism throughout adulthood (e.g. Hong et al., 2016; Renty and Roeyers, 2006; Van Heijst and Geurts, 2015). It may be that a combination of more objective and subjective indicators of a “good” outcome will be most fruitful for future research.
A further complication is that outcomes are often measured at a single point in time. Is the individual working? Does he or she have elevated depressive symptomatology? Yet, a snapshot of adult development will never tell the whole story. Longitudinal studies provide some suggestion that many adults move into and out of “good” outcomes across adulthood (Chan et al., 2017; Howlin et al., 2013; Taylor et al., 2015; Taylor and Mailick, 2014). Psychiatric problems emerge and wane, jobs are procured then lost then procured again, and housing needs and opportunities change depending on a host of factors. Considering outcomes at one point in time cannot tell us about patterns of stability or upward/downward mobility, which are likely more meaningful.
A relevant model for thinking of outcomes in a longitudinal way comes from the debate in educational policy about growth versus proficiency targets in standardized testing. In the educational context, a proficiency target sets a certain threshold (e.g. score of 70) that all students should meet on a post-assessment regardless of where they started, whereas a growth target delineates a certain amount that students should improve from a pre-assessment to post-assessment (American Institutes for Research, 2015). When applied to the context at hand, one could think about a proficiency target being a paid job in the community for at least 20 h a week for everyone, whereas a growth target would focus more on developing a person’s employment skills and prospects a set amount, relative to where they started. As in educational policy, both proficiency and growth approaches have advantages and drawbacks when considering their use in defining adult outcomes (and ultimately the effectiveness of a service or intervention to improve outcomes); yet, these approaches (and the contrast between them) could provide useful frameworks when thinking about how to incorporate a longitudinal approach into the definition of what constitutes a “good” outcome.
In sum, we may need to take a step backward before we can move forward. To accelerate the process of understanding how to effectively and efficiently support individuals with autism throughout adulthood, we should carefully consider which aspects of daily life have the greatest positive impact. We can then use that information to guide our intervention targets. Ideally, consensus could be reached so that studies describing outcomes across various cohorts as well as intervention studies are working from the same rubric when documenting the natural history of “good” outcomes across adulthood and the effects of intervention. Developing consensus on this topic will rely on the input of adults on the autism spectrum, their parent and other important community members as well as the research community. Once consensus has been reached, it will be imperative for researchers to develop standardized tools to measure outcomes that can be reliably used across studies. By laying this important groundwork, we will be able to make more rapid progress in understanding how to support and promote “good” outcomes in ways that are meaningful to those across the autism spectrum and their families.