Digital health should augment (not replace) autism treatment providers

A myriad of digital health technologies have been designed to support autistic individuals. They include software for intervention delivery, augmentative and alternative communication devices, and robots to improve social interaction. Enthusiasm for these technologies in some cases outweighs the evidence. For example, computer programs for teaching social and emotional skills to autistic children have small or no effects (Ramdoss et al., 2012). Our recent randomized trial of a computer-based academic program in autism support classrooms found that, despite teacher’s enthusiasm for the program (Pellecchia, Beidas, et al., 2020), there was no overall benefit to the program (Pellecchia, Marcus, et al., 2020), and it led to reduced use of evidence-based practices (Pellecchia, Beidas, et al., 2020). Similarly, Fletcher-Watson et al. (2016) found that while parents rated a tablet-based intervention highly, it did not improve social communication skills.

Likewise, studies find that virtual reality platforms or robots may teach social and adaptive skills (Boucenna et al., 2014; Goldsmith & LeBlanc, 2004), but those skills don’t generalize (e.g. Whyte et al., 2015), in large part because artificial environments are less effective for teaching generalizable skills than natural settings (e.g. Minjarez et al., 2020). Digitizing treatments reduce opportunities to practice social and communication skills in the real world, and may further isolate individuals who need human connection the most (see Rudd & Beidas, 2020).

In contrast, digital technologies designed to support treatment providers, such as telehealth and progress monitoring platforms, and augmentative and alternative communication devices designed to support autistic people in communicating with others, may be more effective (see Boisvert & Hall, 2014; Ferguson et al., 2019; Kientz et al., 2013; Schlosser & Wendt, 2008; Sutherland et al., 2018) because they help structure therapy sessions, track progress, support client–child communication, and inform treatment rather than replacing it. Telehealth in particular can increase access to treatment and reduce costs (e.g. Lindgren et al., 2016), which is especially important in rural and under-resourced communities, and in the context of our current global pandemic.

Digital platforms such as data collection and progress monitoring apps are designed to reduce burden associated with data collection, a critical component of evidence-based care (Beidas et al., 2019). Digital health technologies can also allow for real-time clinical decision support, to help behavioral health providers know when and how best to intervene (Clausen et al., 2020; Pagoto & Bennett, 2013). They also can be integrated with biosensors to enhance real-time clinical decision support capability using a number of metrics, such as physiological stress, sleep quality, exercise or activity, temperature, and location (Guk et al., 2019; Kuehn, 2016; Nuske et al., 2020). Another real-time clinical decision support technology is augmented reality, as opposed to virtual reality, which allows a provider to conduct their treatment session with augmentations such as easy and real-time access to relevant client or therapy program information (see Son, 2017). For example, a recent randomized control trial showed efficacy in increasing socialization among autistic children using a Google Glass system to promote facial engagement and emotion recognition (Voss et al., 2019).

While these technologies are exciting, three further points should be considered regarding their use in autism intervention. First, most technologies are not rigorously tested before public release. We should hold any digital health technology to the standards set for any autism intervention, including addressing conflicts of interest among these frequently commercialized products (Bottema-Beutel et al., 2021).

Second, there are no best practice guidelines for delivering interventions mediated via digital health technologies. Telehealth is a poignant example. Because of the COVID-19 pandemic, the field has had to grapple with the task of reshaping in-person, therapist-delivered treatment into parent-mediated telehealth with little best practice guidance. Providers have been tasked with teaching families how to engage in telehealth services without training (Vismara et al., 2012). Research is sorely needed on how to improve the telehealth experience for autistic children and their families, and on which children benefit most from telehealth services. Indeed, provider and parent training is important for incorporating digital health technologies in treatment (Baxter et al., 2012).

Third, we must address the disparities in access to digital health technologies. Although technologies to augment autism interventions were once an attractive new option, the COVID-19 pandemic imposed these technologies as the new reality. Through the pandemic we have learnt much about issues of digital accessibility. Digital health technologies necessitate, in many circumstances, access to high-speed Internet and a technical skill that autistic individuals and their families may not have. The well-documented digital divide may exacerbate disparities, if the quality of services received is directly related to the quality of the technology that the family can afford (Ameis et al., 2020). This is important as many organizations, companies, and schools have opted to continue to offer telehealth services and distance learning to their clients and students, respectively, beyond the pandemic.

Digital health technologies offer some exciting new avenues for improving the delivery, monitoring, and coordination of autism programs. We urge the field to rigorously test these technologies prior to public release, work toward best practice guidelines for digital health technology implementation, and consider methods to promote equitable access to digital health technologies that augment rather than replace autism treatment providers.