When signing into Amazon to purchase a book, the algorithm automatically evaluates the choice and artificial intelligence lists other books that may be of interest. The SoundHound app is able to provide information about a song, and helps the user to find other music that they may enjoy. Technology has moved from entertainment applications to streamlining healthcare. There are now algorithms for mental health, digital health technologies, that enable chatbots or other apps to respond to human emotion and pain, and provide AI generated solutions.

The evolving popularity and use of these programs have grown tremendously during the pandemic, and the shortage of psychiatrists before the pandemic is now at crisis levels given the growing mental health needs in the United States and other countries. A study in Canada notes that only half of Canadians receive mental health treatment that is considered adequate. The World Health Organization acknowledges that almost half of the entire world population lives in nations with less than one psychiatrist per 100,000 people. As mental health apps and other tools grow in quantity, and more people are using them, it is imperative to determine if these interventions are effective.

A series of research articles notes that apps like chatboxes are increasingly being used, and may have some advantage over human mental health professionals. The chatbox, as noted in a research article, is not distracted by a busy schedule or having to complete electronic medical records that human clinicians are burdened with. In addition, a chatbox can use machine learning and artificial intelligence to remember everything that a patient has said, and can become smarter with each interaction, tapering its responses to become more effective and focused on the patient at hand. These chatbots can even provide cognitive behavioral therapy, a highly respected and evidence-based treatment developed by the late Aaron Beck at the University of Pennsylvania, and proven to be effective to treat anxiety and depression. On the other hand, many psychiatrists and therapists, who are aware of the effectiveness of cognitive behavioral therapy, may find using this difficult if they have not been trained in this, and may also find it difficult to avoid “winging it” and using non-evidence-based therapies because that is how they have been trained. In fact, many therapists focus on using analytic tools such as psychoanalytic frameworks that have never been sufficiently tested for effectiveness. The University of Scranton notes that there are now over 500 different types of psychotherapies, and it is clearly not possible to evaluate them all. The chatbox or mental health app doesn’t have inherent cognitive and emotional biases as humans do, and will not become distracted by fatigue or stress, as many humans have. There is evidence that mental health apps, which number over 315,000, may be effective if they are supported by ongoing personal visits with a trained clinician. Therefore, it is not necessary to throw out the baby with the bath water.

A research article by David Gratzer, MD, and David Goldbloom, MD, at the University of Toronto, describes this therapist-supported treatment as being effective for a number of psychiatric conditions, including social anxiety, panic disorder, post-traumatic stress disorder, generalized anxiety disorder and major depression. Although more research is needed, it appears to be that the collaborative efforts of human with machine may be more effective than machine alone. More research is needed to validate this claim.

Dr. John Touros at Harvard Medical School, who was interviewed for this column, notes in a published report that when patients describe suicidal thoughts to an app, it may provide answers that are not helpful. Using the app is also incongruent with the number of times it has been downloaded. Touros, et al., writes that in the case of PTSD Coach App, the Veteran’s Administration found that this app was downloaded over 150,000 times, but the day after downloading it, only 14% of patients had actually used it. This is similar to Amazon book downloads, where an individual might purchase a book but never read it.

In discussing the evolving role of AI in mental health, the prime minister of Canada, Justin Trudeau, stated that “if we can use apps to order dinner … we can use technology to keep each other healthy.” In an article by David Gratzer and John Torous, et al., digital health is noted to improve access to care, because it is not bound to the geographic limitations of patient or clinician locality. It may also improve adherence to treatment and it certainly is more convenient than arranging an appointment with an in-person psychiatrist or therapist. Telepsychiatry, which has become more widely adopted during the pandemic, is unfortunately less studied for patients who have severe mental health issues, such as schizophrenia, and who are off the grid because of poor socioeconomic status and inconsistent living situations. Even the pandemic, which highlighted the mental health needs of all U.S. citizens, still had a brick wall of discovery for severe mental illness, and those with schizophrenia who were actively suffering were somehow forgotten even during this healthcare crisis.

Digital health, at least theoretically, provides equal access to all.

Digital literacy, and comfort with digital tools, also impacts whether someone will use virtual tools or avoid them. Privacy is also an issue. Ideally, these tools will be connected or compatible with electronic health records, which will obviate clinicians having to review them in addition to all of the other work that is required in a 15-minute medication check. Optimally, with the transition to value-based care and proactive measures, the data from these online and virtual apps will be able to automatically share and analyze this information for clinician review. Like any therapeutic modality, it will be important to discern the optimal temporal use of these interventions. Research will be necessary to find the “sweet spot” of clinical efficacy, and to understand the individual thresholds for patient efficacy. It may be that there is a downward slope of efficacy if these tools are used to little, or too much.

Anecdotally, some clinicians have adopted telepsychiatry for their safety, and the safety of their patients, throughout the pandemic. There are some clinicians who may, on the other hand, be cynical about telepsychiatry and are also unfamiliar or uncomfortable with this technology. Patient demand may stimulate many providers to become more efficient using telehealth.

The demand for evidence-based mental health care would require an additional 4 million more practitioners to meet this need, and digital health technologies, that include apps, chatbots, virtual reality and social media, can clearly augment the supply of interventions for patients in need. These are interventions that surpass the electronic health-records or the virtual psychiatry being conducted over Zoom or other meeting technologies

A Feb. 13, 2012 article in the New York Times by Benedict Carey, evaluates the immediacy of utilizing digital mental health technologies. Downloading an app onto a smart phone makes therapy available at a retail store, while waiting in line for groceries at checkout, or before finishing chores at home or work.

Being ubiquitous doesn’t automatically translate into being effective. In an interview with John Torous, MD, the director for digital psychiatry at Harvard, he states that there are more than 700 online apps available. However, despite the abundance, and the growing number, none of these apps have completed truly evidence-based studies. “Not one app is better than another,” Torous notes. “Even Pear Therapeutics, which has FDA approval for some of its products, has weak data supporting its digital prescription therapeutics, and is not scalable to a large group of people.” He continues that “the future is hopeful that eventually data will exist that can demonstrate whether an app is truly effective.”

He refers me to ICER, the Institute for Clinical and Economic Review, which has formally studied digital prescription therapeutics and their effectiveness. Digital prescription therapeutics is a brand new type of prescription. It is software built to help address addictions or PTSD. None of the studies of these apps, according to the ICER, “has any data on the outcomes that matter to patients” and there was the complete absence of any peer-reviewed scholarly academicians that are required to publish articles in many professional journals. One study noted that the key health outcomes necessary for measurement, such as being able to return to work, avoiding the ER and hospitalizations and enjoying an improved relationship with family, were not measured and completely lacking.

Nonetheless, there is hope that such data will eventually be measured, and will be able to benchmark apps for efficacy and outcome studies. The internet has made information and virtual tools exponentially abundant and, once the research data can discern which digital tools are most effective for mental health, access and enrollment can be as simple as entering an email address or signing up with a credit card. However, this is going to be a challenging task to complete. The app CALM, which provides, among other benefits, an audio library of soothing sounds for sleep — such as rain or water sounds — is not unique. It is possible to find these sounds for free on the internet, so why do consumers pay for this? Perhaps they are unaware how redundant the features of CALM really are, and are possibly recruited through television advertising. However, ethical and moral healthcare should ultimately be provided by apps that are unique in features, and that have been proven to provide sustainable and repeatable outcome studies for their effectiveness, ease of use, and demonstration that any untoward possible side effects can be knowingly avoided. The first creed in medical school is “first do no harm.” These apps will have to jump that hurdle and, at the same time, consistently help improve the areas that matter most to people in their lives.

— — —

This is part one of a three-part series.