Cognitive Technology for Addiction-It's Time for Action

The 2015 US federal budget includes $27 billion targeted for drug control. This amount does not take into account about another $25 billion in other related costs. Think about this: 1.5 million people are arrested each year; approximately 345,000 people infected by HIV due to needle sharing; and approximately 2,220,300 people in prison or local jails.

The statistics demonstrate only the economic costs. Other costs include the damage done to families, children, and marriages, the loss of jobs, lost productivity, accidents, the list goes on and on. Addiction is very expensive indeed.

There are a wide variety of addicts including approximately 17 million people in the US with alcohol use disorder and many millions more addicted to a variety of drug classes. More broadly, addiction also includes about 42 million smokers.

Relapse is the Hallmark of Addiction

One of the hallmarks of drug addiction is the relapse rate. For some reason, our medical delivery system treats addiction like a hernia repair: intervene and it’s fixed. But it isn’t. Addiction-related recidivism is high for a number of reasons including, of course, the addictive nature of substances. But recidivism is also associated with a lack of insurance coverage as well as an absolute deficit of trained professionals capable of helping with addiction.

Other barriers include the personal reasons for people to not seek treatment such as stigma and embarrassment, as well as a desire to try to tackle this complex behavior change on their own.

Cognitive Technologies Offer a Proven Option

Numerous studies have been done on how to stretch precious human resources using cognitive technologies. One, written by John Cunningham et al, entitled The Use of Emerging Technologies in Alcohol Treatment described a number of internet-based interventions and divided them into 3 groups:

1. Screening Tools
2. Cognitive Behavior Treatment Programs
3. Emerging Technologies

There are several screening tools available. Many are included by the Substance Abuse and Mental Health Services Administration, a government agency dedicated to improving health.

These screening tools can be used in a variety of age groups and include scoring processes. Many have demonstrated efficacy in not only improving screening but also in actually changing drinking behavior.

Internet-based cognitive behavior treatment programs also have proven efficacy for addiction treatment. These include functions such as text messaging as well as online teaching and counseling.

Online support groups also can be used to assist with alcoholism. Alcoholics Anonymous and Moderation Management have proven efficacy in a variety of populations.

Solutions Must Consider How the Brain Works

System 1 thinking is efficient, requires little active thought process, is automatic, rapid, contextualized, intuitive, independent of working memory, and is basically a series of habits or reflexes. This is important because it is estimated that over 95% of our actions do not involve the conscious reasoning or System 2 thinking.

System 2 thinking (slow brain) is slow, controlled, requires high effort, is rule based, and logical, but is not used much by humans for day-to-day activity as it just takes too much energy and concentration.

It is important that any technology solution becomes so easy to use that it ties into System 1 or “fast brain” thinking. As a quick review:

Most digital approaches fail in one major way: they require the participant to actively start the engagement and maintain it. This requires System 2 thinking and humans just don’t like to think! We know that the easier the process, the more people will flock to it. Think about how other masters of the digital world, such as Amazon, have used System 1 thinking. How could we emulate their amazing results with healthcare?

For one, many populat internet sites use natural language understanding to find the item you are searching for; even if you do not know the exact word to find the article. Amazon also constantly suggests additional items that others usually purchase together under the heading “Customers Who Bought This Item Also Bought.” They have created a “Buy now with 1-click” button and are constantly asking you to consider re-purchasing previous purchases as well as asking for feedback on the items you have purchased. All of this helps to drive outcome.

Other popular digital companies, such as Uber, have utilized your actual location (using GPS tracking) as a tool to engage with people with relevant services. Still others, such as Facebook and LinkedIn, have used a social cognitive approach to rapidly grow their companies to massive levels because of the innate human desire to join groups and keep in touch—and they have made it simple by using massive analytic engines in the background.

All of these companies utilize System 1 thinking processes. They make it easy and use the subconscious, emotional, and habit centers of the brain to allow for a more engaging interaction. And these approaches have changed our behavior in a variety of ways.

David Gustafson, a professor at the University of Wisconsin, has created, and studied in a randomized fashion, a smartphone application to support recovery from alcoholism and published extensively on this topic. In his article, he describes a mobile technology application called Alcohol-Comprehensive Health Enhancement Support System, (A-CHESS), which runs on a smartphone. He utilized a “self-determination theory” as the theoretical basis that “posits that meeting 3 needs contributes to an individual’s adaptive functioning.” These 3 needs are: being perceived as competent, feeling related to others, and feeling internally motivated and not coerced in one’s actions.

The A-CHESS application supplies both static content (audio-guided relaxation) as well as interaction features. It also included a warning function using the GPS function of the smartphone: if a patient traveled to a location of a frequently visited bar, it would ask the patient if he really wanted to be there. The application also included a “panic button” that could contact 2 people who acted as accountability partners. Trained counselors monitored the data.

This study demonstrated high participation. It also determined that patients in the A-CHESS group reported significantly fewer “risky drinking days” greater abstinence as well as higher scores for “abstinence confidence.” Although there were some limitations and a need for further studies, this “state of the art” intervention demonstrated that a digital, infinitely scalable approach is worthy of consideration. Of interest is that this research has lead to a commercial product from a company named CHESS Mobile Health, Inc.

Cognitive Technologies Come to Medicine

Although A-CHESS is a remarkable development in the treatment of drug addiction, I propose an even more remarkable approach to this illness; the addition of natural language understanding in the form of a Virtual Health Assistant (VHA).

A VHA is several generations beyond the typical smartphone “app.” It is the result of the culmination of advanced smartphone technology: artificial intelligence (AI), natural language understanding (NLU), and cloud-based computing combined to create a centerpiece of what Deloitte has termed "cognitive technology."

In their white paper, the authors outlined the inevitable need for this approach in healthcare. A VHA can emulate humans. It brings the power of the “predicate” technologies to an individual person in a here-to-for unprecedented manner. And, the big difference is that a VHA can talk!

Because of AI and NLU, a VHA can literally enter into a conversation with you. The NLU function can understand regular human speech, not just as “voice recognition” but beyond that, it can understand the intent of the conversation and act on that understanding. A VHA acts in a way that a human might respond to a statement or question, hence the term "human emulation."

Building on the A-CHESS approach, a VHA can use smartphone functions such as the GPS. But it can also literally take over functions like dialing 911 for an emergency. It can also act sort of like an “air traffic controller” for all of the other apps used by that person; launching them as needed and recording the data using conversation as the tool to collect the data. A VHA can provide encouragement and assist in motivation by providing adult learning principles and content. Programmed using motivational interviewing techniques, it can engage with people at a scale never dreamed of using professionals.

A VHA can also “sense” moods from conversation and react accordingly. For instance, if a VHA asks a simple question such as: “Good morning, John, how are you today?” and John responds, “I am sad.” The VHA can ask the pertinent depression screening tool and score it.

Additionally, a VHA can answer tens of thousands of different questions within a domain and if desired, launch a structured teaching program and determine when a person has mastered the content of a specific course of material; something highly desirable in a field such as addiction where knowledge gaps are common.

A VHA can also utilize the proven psychological approaches to change behavior and deflect a person’s actions using proven triggers to more positive behaviors. It can also be used to create community by connecting groups, in an anonymous fashion, others who can act to encourage accountability, similar to the website Patients Like Me.

Not Science Fiction

Although this technology seems a bit “out there,” it isn’t. The core technology has been around for many years in other industries and is highly developed and proven. The behavioral approaches have also been proven using a human-based approach as well as less sophisticated technology such as “apps.” By combining NLU and AI, we can leapfrog the older technologies much like the Blu-ray DVD technology has made the VHS obsolete.

In fact, an addiction VHA is already in development. A small company called Life Recovery Solutions (LRS) is in the development phase of a virtual addiction coach using NLU and AI. LRS' goal is to create behavioral modification tools using NLU and AI.

Addiction is a national scourge. We have neglected it and focused on a “war” approach that has not worked. It is time to focus on helping those addicted using a proven technology and proven behavioral approaches. A virtual addiction coach can dramatically change our approach to recidivism and can be made available in the pocket of virtually every person in need in the US within a few months.

But, for this technology to be widely used large healthcare organizations, as well as our justice system, must step up to the plate and start to offer this service. The need is great, the technology is ready; who will be the first to adopt it?