The 2022 release of ChatGPT pretty much made artificial intelligence (AI) a household name. You can barely turn on the television, clean out your email inbox, or scroll social media without coming across something AI-related.
Of course, the healthcare space often lags behind the technology curve, and we’ve certainly seen that with AI—though certain specialties, like imaging, have embraced it more readily than others. (Not all that surprising, considering that it’s much easier to train AI systems on images versus human language.)
Still, while behavioral health is largely conversation-based—and thus, more difficult to parse out with AI—this technology has made significant inroads in our industry in two ways:
- Digital therapeutics (DTx), and
- Augmented intelligence (AIx).
Let’s walk through these two vastly different applications of AI technology in behavioral health—and why providers should approach each with both curiosity and caution.
What are Digital Therapeutics?
Perhaps the most obvious difference between digital therapeutics and augmented intelligence is the end user of each tool.
DTx tools are typically tailored to healthcare consumers. In fact, the term “digital therapeutic” specifically refers to software tools that, similar to a pharmaceutical, have been certified by the Food and Drug Administration. According to the Digital Therapeutics Alliance, there are currently 19 FDA-certified software products—eight of which target behavioral health problems.
The Rise of DTx in Behavioral Health
That said, there are tons of consumer-facing tools out there that have not been certified by the FDA—like the estimated 10,000 to 20,000 behavioral health tools available in the Apple and Google app stores. While “digital therapeutics” technically require FDA approval—and must be prescribed to patients by a qualified healthcare provider—we will use this term more liberally for the purposes of this article, encompassing any consumer-facing digital health tool.
Due in part to COVID-19, we’ve seen an explosion of DTx products over the last few years. Like telehealth, they existed prior to the pandemic, but COVID protocols amplified the need for remote therapeutic interventions.
In general health, DTx tools often target patients with chronic disease conditions such as diabetes and hypertension. In behavioral health, targeted conditions run the gamut of diagnoses—but many DTx tools are designed to treat more acute problems such as anxiety and depression, or to provide wellness support by teaching things like mindfulness and stress management.
The Major Limitations of DTx Tools
While these tools can be useful in some cases, it’s important to understand the following problems and limitations:
- Limited Research Base: According to the Agency for Healthcare Research and Quality (AHRQ), “Thousands of behavioral health apps are currently available in commercial app stores, but it’s been estimated that only 3 to 4 percent are evidence based. While some of these apps have been extensively studied and shown to be efficacious, the vast majority (an estimated 98 percent) have not been systematically evaluated.”
- Poor Engagement: One of the major challenges with behavioral health DTx tools is user engagement. As this JMIR study notes, “Engagement with most mental health apps is abysmal—estimates suggest that most publicly available mental health apps for depression and anxiety have zero or near-zero active users…Estimates suggest that approximately 4% of users who download a mental health app continue using it after 15 days, and 3% continue after 30 days.”
- Accessibility and Usability Issues: Not all users have equal access to digital tools due to disparities in digital literacy, internet access, and device availability. Furthermore, these tools may not be designed with all user needs in mind—particularly those of older adults or individuals with severe mental health conditions. This can create barriers to access and effective use of DTx tools.
- Data Privacy and Security Concerns: Users often worry about the privacy and security of their personal data when using digital therapeutic tools. Given the sensitive nature of behavioral health data, any breach or misuse of information can have serious repercussions. This concern can deter individuals from fully engaging with digital tools.
- Lack of Ongoing Support: Unlike traditional therapy, digital therapeutic tools may not offer continuous support—which, in turn, leads to discontinued use. Users may struggle with the lack of human interaction and the absence of human feedback, which are often crucial to the success of behavioral health interventions.
A huge part of the problem with DTx tools is that they have been marketed as a replacement for trained clinicians—not a tool to support the care they deliver. While leveraging DTx tools as a substitute for provider-delivered care made some sense at the height of COVID, it also led to a precipitous decline in long-term consumer usage.
This might all look different if DTx tools were positioned as supplemental to the care process. For example, a clinician might assign a patient to use a particular app that transmits information back to the clinician via the EHR. This “clinician in the loop” model combines the power of both face-to-face and digital interventions—ultimately leading to higher engagement and potentially better outcomes.
What is Augmented Intelligence?
According to the Institute of Electrical and Electronics Engineers (IEEE), augmented intelligence is defined as a “subsection of AI machine learning developed to enhance human intelligence rather than operate independently of or outright replace it. It’s designed to do so by improving human decision-making and, by extension, actions taken in response to improved decisions.”
The American Medical Association (AMA) has adopted a similar view on augmented intelligence, defining it as “a conceptualization of artificial intelligence that focuses on AI’s assistive role, emphasizing that its design enhances human intelligence rather than replaces it.”
So, the defining characteristic of augmented versus artificial intelligence is its assistive role to clinicians. For the most part, any healthcare provider leveraging artificial intelligence is doing so in an augmented intelligence capacity.
Augmented intelligence tools like Eleos are intentionally designed to help providers do their jobs faster, better, and easier. They are essentially enhanced clinical decision support systems that give clinicians information they can use to make better decisions about the care they are delivering. To be clear, these systems are not making decisions for clinicians.
Digital Therapeutics vs. Augmented Intelligence
| Digital Therapeutics | Augmented Intelligence | |
|---|---|---|
| Definition | Software-based interventions to treat, manage, or prevent mental health disorders. | AI systems that enhance clinical decision-making, clinical documentation, and patient care by providing insights and recommendations. |
| Purpose | To provide standalone or adjunct treatment for specific behavioral health conditions. | To assist healthcare providers in making more informed decisions by analyzing data. |
| Usage | Used directly by patients, often as part of a prescribed treatment plan. | Used by clinicians to support treatment, problem identification, care planning, and patient monitoring. |
| Regulation | Often requires regulatory approval (e.g., by the FDA) as a medical device. | Generally not subject to the same regulatory scrutiny; viewed as a clinical decision support tool. |
| Examples | Apps for cognitive behavioral therapy (CBT), medication adherence tools | AI-driven tools such as Natural Language Processing (NLP) to generate clinical insights and accelerate documentation. |
| Interactivity | Highly interactive, involving patient engagement and behavior tracking. | Less interactive, focusing on data processing and information output for clinicians. |
| Impact on Patient Care | Provides direct therapeutic benefits to patients, often without clinician involvement. | Enhances a clinician’s ability to deliver personalized care, but does not replace their clinical judgment. |
| Data Utilization | Uses patient-reported data and outcomes to tailor treatment recommendations. | 1. Uses ambient listening to provide information/insights to clinicians in near real-time. 2. Provides improved ease of use for mobile workers to improve and expedite documentation. |
| Example Technologies | Pear Therapeutics’ reSET, Akili Interactive’s EndeavorRx | Eleos, Limbic, Lyssn, Bells AI |
Why Should Behavioral Health Providers Embrace AI?
Unfortunately, the term “AI” often evokes memories of science fiction movies where robots have run amok, taking over jobs and lives and even destroying humanity as we know it. Even when these irrational, Hollywood-fueled fears are set aside, clinicians often ask whether AI will take their jobs. So, the fear of replacement is still very real.
The short answer to this question is no—AI will not take your job as a behavioral health provider, because:
- In healthcare, the vast majority of AI falls into the augmented intelligence category, meaning it is designed not to replace clinician judgment, but rather, to enhance it.
- There aren’t enough behavioral health providers to meet the current need—and that deficit is projected to get worse. It’s not a great situation for clients, but it does offer a promise of job security.
But here’s the thing: while it’s true that AI will not take your job, it’s also true that someone using AI will take your job. What I mean by that is augmented intelligence is the future for all of healthcare.
This is not because it is the “fad du jour”; it’s because it is literally impossible for any healthcare provider to know all that there is to know about their discipline. Healthcare knowledge, like all knowledge, has a half-life (i.e., the amount of time it takes before half of all that is known about a particular subject is obsolete).
For psychology, that time is about seven years. For general medicine, it’s about 73 days.
The point is, no one can stay current on all that is knowable via existing educational mechanisms like CEUs and conferences. Augmented intelligence can help with this. In addition to supporting faster session documentation, specialized AI tools act as another set of “ears,” similar to having a co-therapist.
In the future, I believe AI tools will use what they “hear” to advise clinicians about current, relevant research—thus addressing the “half-life” problem. We are at the beginning of a new era of evidence-informed care, and clinicians of all disciplines should start learning how to integrate AI systems into their workflows now—because this is the way care will be delivered going forward.
Ready to take your organization into the future with purpose-built behavioral health AI? Request a demo of the Eleos platform here.
References
- Dorsey, E. R., & Topol, E. J. (2020). State of Telehealth. The New England Journal of Medicine, 375(2), 154-161 State of Telehealth – PubMed (nih.gov).
- FDA. (2021). Digital Health Software Precertification (Pre-Cert) Program. U.S. Food & Drug Administration. Retrieved from FDA Digital Health Software Precertification (Pre-Cert) Pilot Program | FDA
- Topol, E. (2019). High-performance medicine: the convergence of human and artificial intelligence. Nature Medicine, 25(1), 44-56 High-performance medicine: the convergence of human and artificial intelligence – PubMed (nih.gov).
- High-performance medicine: the convergence of human and artificial intelligence | Nature Medicine
- What is Augmented Intelligence? – IEEE Digital Reality
- AI Should Augment Human Intelligence, Not Replace It (hbr.org)
- Augmented Intelligence: Enhancing Human Decision Making | SpringerLink
- Indian Journal of Urology (lww.com)
- Making Policy on Augmented Intelligence in Health Care msoc1-1902_2.pdf (ama-assn.org)
- Is It Ethical to Use Prognostic Estimates from Machine Learning to Treat Psychosis? cscm1-1809_0.pdf (ama-assn.org)
- An Intelligence in Our Image: The Risks of Bias and Errors in Artificial Intelligence | RAND
- Hybrid-augmented intelligence: collaboration and cognition | Frontiers of Information Technology & Electronic Engineering (springer.com)
- 6 Digital Therapeutics Challenges & How to Overcome Them (binariks.com)
- The ascent of digital therapies (apa.org)
- Digital mental health: challenges and next steps | BMJ Mental Health
- Role of digital therapeutics and the changing future of healthcare – PMC (nih.gov)
- Digital therapeutics (DTx) for disease management | McKinsey
- the-health-benefits-and-business-potential-of-digital-therapeutics.pdf (mckinsey.com)
- Mapping the regulatory landscape for artificial intelligence in health within the European Union | npj Digital Medicine (nature.com)
- Behavioral Health Apps in Primary Care | The Academy (ahrq.gov)
- User Experience, Engagement, and Popularity in Mental Health Apps: Secondary Analysis of App Analytics and Expert App Reviews – PMC (nih.gov)
- Clinical Evaluation of Digital Therapeutics: Present and Future – PMC (nih.gov)
- Enhancing Integrated Behavioral Health Care with Digital Therapeutics Bing Videos
- A systematic review of digital and face-to-face cognitive behavioral therapy for depression | npj Digital Medicine (nature.com)
- Efficacy of stand-alone digital mental health applications for anxiety and depression: A meta-analysis of randomized controlled trials – ScienceDirect
- The Rise of Prescription Digital Therapeutics in Behavioral Health – PubMed (nih.gov)
- Frontiers | Digital therapeutics for mental health: Is attrition the Achilles heel? (frontiersin.org)
- Ethical considerations of digital therapeutics for mental health – ScienceDirect
- What is Digital Health? Digital mental health technologies (apa.org)
- Home – Digital Therapeutics Alliance (dtxalliance.org)
- International Organization for Standardization (ISO) Digital Therapeutic Definition—June 2023 DTA_FS_ISO-Definition.pdf (dtxalliance.org)
- Digital Therapeutics Definition and Core Principles DTA_DTx-Definition-and-Core-Principles.pdf (dtxalliance.org)
- AI is changing every aspect of psychology. Here’s what to watch for (apa.org)
- Digital therapeutics from bench to bedside | npj Digital Medicine (nature.com)
- The Expanding Role of Digital Therapeutics in Mental Health (psychiatrictimes.com)
- Digital therapeutics and mental health apps (apa.org)
- Prescription Digital Therapeutics For Mental Health: What It Means For Doctors And Their Patients (forbes.com)
- Role of digital therapeutics and the changing future of healthcare JFMPC-9-2207.pdf (nih.gov)
- Mapping the regulatory landscape for artificial intelligence in health within the European Union | npj Digital Medicine (nature.com)
- Determinants of and barriers to adoption of digital therapeutics for mental health at scale in the NHS | BMJ Innovations
- Do mental health mobile apps work: evidence and recommendations for designing high-efficacy mental health mobile apps – PMC (nih.gov)
- Behavioral Health Apps in Primary Care | The Academy (ahrq.gov)
- User Experience, Engagement, and Popularity in Mental Health Apps: Secondary Analysis of App Analytics and Expert App Reviews – PMC (nih.gov)
- Barriers to and Facilitators of User Engagement With Digital Mental Health Interventions: Systematic Review – PMC (nih.gov)
- FDA Permits Marketing of First Game-Based Digital Therapeutic to Improve Attention Function in Children with ADHD | FDA
- The State of Digital Therapeutics and FDA Regulations (hattrick-it.com)
- Methods for Navigating the Mobile Mental Health App Landscape for Clinical Use – PMC (nih.gov)
- FDA Roundup: April 2, 2024 | FDA
- Computerized Behavioral Therapy Device for Depression Disorders Product Classification (fda.gov)
- FDA Approval and Mental Health Apps (verywellmind.com)
- The FDA has approved the first app to help treat depression : NPR
- Digital Therapeutics for Management and Treatment in Behavioral Health https://store.samhsa.gov/sites/default/files/pep23-06-00-001.pdf
- Digital mental health: challenges and next steps | BMJ Mental Health
- Digital mental health apps and the therapeutic alliance: initial review S2056472418000868jrv 1..5 (harvard.edu)
- Augmented intelligence in medicine | American Medical Association (ama-assn.org)
