Wearing recovery on one`s sleeve

While there have been many advances in the treatment of addiction, relapse remains a common problem.1,2 Research indicates that 25 to 50% of people who have undergone addiction treatment moved back and forth between periods of abstinence and heavy drinking or drug use.3,4 Not surprisingly, rates of relapse increase consistently as more time passes since treatment.

Although most field experts consider addiction to be a chronic disease, there are often limited provisions for clients who have completed treatment to manage their addictive disorders through ongoing monitoring or “checkups,” which are seen more commonly with other chronic disorders.5 Recent studies indicate that the key to preventing addiction relapse often lies in experiences and events in the hours and even minutes leading up to the onset of an episode of returning to substance use.6 Some of the most common reasons for first relapses are negative emotions, social pressure, and cravings to use a substance.7 One explanation for these relationships is that people return to substance use as a way to escape from distress or as a means of coping with stress and negative emotions.8

Importantly, each of these stressors that contribute toward relapse reveals itself in physiologic forms within the body—all of which can be detected via wearable biomonitoring technology available on the market today.

“We know that stress is the most accurate predictor of relapse, and relapse is often preceded by stressful episodes characterized by strong negative moods,” says Timothy Baker, PhD, professor of psychology at the University of Wisconsin-Madison. “This suggests that autonomic nervous or somatic system activity associated with stress may be used to detect stressful episodes and make people aware of their state of mind. One reason this is important is that laboratory research shows that individuals may actually be unaware of when they are becoming stressed or experiencing negative moods, so they don't always know to seek out the help they may need to maintain a successful recovery program.”

Our research team at the University of Wisconsin has begun testing the potential of how body monitoring instruments can be used to prevent relapse. This research is part of a larger initiative called “Innovations for Recovery” that is focused on exploring how technology can improve outcomes in addiction treatment.9 The idea we pursue here is the feasibility of using physiologic monitoring to detect stress and to trigger data services such as a text message or phone call to the user asking if he/she needs support.

Possible applications

Bret R. Shaw, PhD
Consider the following scenario of how a wearable biomonitoring system might help prevent relapse. A woman in recovery from alcoholism is waiting at home for her husband to return from a fishing trip. He is often late and has already promised to be home in time for a dinner engagement they had made with friends weeks earlier. As the time for their dinner approaches, the woman feels increasingly anxious and tempted to open a bottle of wine that her husband had received over the holidays. Her wearable biomonitoring device detects her distress and sends a message to her mobile phone asking if she needs any support. Likewise, the sudden increase in her stress levels triggers text messages to the mobile phones of people in her support network, such as her peer sponsor and a therapist. They immediately phone to make sure she is managing her stress appropriately (i.e., not drinking).

The most appropriate application we could identify to pursue this research aim of exploring how such a system might work was the BodyMedia SenseWear body monitoring system (https://www.bodymedia.com), a monitor that collects the physiologic information needed to measure arousal and that incorporates integrated wireless data transmission and Web-based interfaces for interpreting results. Both co-authors of this article did our own appraisal of this device and its possible efficacy as part of a comprehensive addiction relapse prevention system.

We both found the armband surprisingly comfortable and easy to wear. After the first hour or so, both of us forgot we even had it on unless someone reminded us or we bumped into something. Also, both the software and the wireless port that transmits data between the armband and the computer were easy to install and configure.

The SenseWear armband combines multiple sensors for motion, skin temperature, galvanic skin response, and heat flux. Both of us wore the armband over several extended periods that encompassed stumbling into stressful situations as well as knowingly entering such situations. We both found that galvanic skin response and heat flux were most clearly associated with our own subjective experience of stress.

Andrew was sufficiently committed to our scientific inquiry to hold off on resolving an ongoing disagreement with his live-in girlfriend until he was wearing the armband. It had been a hectic couple of days for both of them, and Andrew could sense that his girlfriend's dissatisfaction with the lack of quality time they had been spending together was building. After they returned home from a rare night out with friends, Andrew went straight to his desk to do a little work, which was not what his girlfriend was hoping he would do. Andrew's galvanic skin response started to climb at 3:20 a.m., when his girlfriend explicitly shared her disappointment with him and a confrontation ensued. As the argument came to a hilt at 3:32, Andrew's heat flux spiked, along with his galvanic skin response.

Andrew Isham
To provide a contrast, we examined physiologic data collected during a vigorous game of basketball that Andrew played later the same week. Again, Andrew's galvanic skin response and heat flux increased and stayed elevated throughout the game. But whereas during the argument with his girlfriend his acceleration stayed at or slightly above baseline levels, his acceleration levels during the basketball game increased substantially because of his spirited activity on the court.

This contrast raises a crucial issue in using data such as this to trigger needed supports in the context of relapse management. One of the challenges in detecting stress that could lead to relapse lies in avoiding “false positives—interpreting a situation as stressful when the physiological signals simply reflect healthy arousal from activity such as exercise. To distinguish between vigorous physical activity that does not warrant any follow-up and negative emotional arousal that could lead to relapse, it might be necessary to map out multiple data points. For example, one might develop an algorithm that triggers queries to a person in recovery only when certain indicators rise in the absence of a corresponding increase in physical movement.

It also should be noted that the purpose of monitoring physiological stressors would not be to identify risk situations flawlessly. Rather, the purpose would be to serve as a relatively sensitive risk detection device prompting the individual to monitor his/her ongoing situation and take coping actions if needed. The intent would be to help the person to use cognitive control processes and not to react to stressors automatically with maladaptive coping methods.

Bret's most notable experience while wearing the SenseWear armband occurred on a Sunday afternoon while he was at home working on an article, with the kids at home and his wife at the office. His 12-year-old stepdaughter, Mahala, asked if he could drop her off at the cinema. He gave her mother a call to make sure it was okay. Her mother asked who she was going with, and when Bret asked Mahala, she shouted back, “Sergio.” That would have made this her first date to a movie with a boy—something for which Bret and his wife were not fully ready. Her mother said she could go if other friends went, too. Mahala got on the phone to invite a larger group of friends, and later in the afternoon Bret drove her to the movies as promised.

When Bret and Mahala arrived at the cinema, the tall, handsome Sergio stepped out of the theater doors and waved; there were no other friends to be found. “You can go on,” Mahala said. “My friends will come.” Bret reminded Mahala of her mother's instructions that he was not to leave her without her girlfriends present—a comment that was not particularly well-received. The start of the movie approached, and the time continued to pass without her friends arriving. Bret could feel his anxiety rising but didn't really think about the armband he was wearing and what it might be recording as a physiologic indicator of his inner state of affairs (a combination of guilt and resentment about feeling like a controlling person). Finally, Mahala's friends arrived, and everyone hurried into the theater to catch the movie.

That night, Bret took off the armband and transmitted his physiologic data to his laptop. There was not a lot of variation occurring in his measurements until right at 4:20 p.m. when he and Mahala had arrived at the theater, at which time his galvanic skin response and especially his heat flux had risen dramatically above all his other physiologic indicators. Bret's wife and Mahala had a good laugh looking at this phenomenon, which they dubbed “the Sergio effect.”

Interestingly, Bret did have a few glasses of wine that afternoon after coming home from the theater without really thinking much about it. He had made previous well-intended commitments to himself to cut back on drinking of this sort. Is it possible that a reminder that he was in fact stressed, along with encouragement not to drink, might have had a positive effect on this behavior? Bret certainly doesn't think it would have hurt.

Comparisons with biofeedback

The capabilities that could be enabled by the BodyMedia SenseWear body monitoring system are in some ways similar to basic biofeedback technology conceived in the 1960s. Both are intended to make people aware of their physiologic stress, which can empower them to learn to respond adaptively rather than in ways that might set back their personal goals.

However, there also are significant differences between traditional biofeedback technologies and the capabilities described in the system we are discussing here. Perhaps most importantly, biofeedback is not an active form of feedback—nothing is being done to the person when he/she is hooked up to biofeedback equipment. The system we are talking about in this article detects stress in order to trigger needed services such as automatic text messages with concrete recommendations designed to prevent relapse, or communications from supportive others such as a peer sponsor or therapist.

While implementing these more sophisticated technologies could increase the likelihood of providing supports when needed, their limitations also must be recognized. Currently used instruments are typically not very specific—they are accurate for detecting general emotional arousal rather than specific emotional states such as anxiety or depression. In other words, an increase in body temperature can be associated with emotional or physiologic arousal but does not indicate whether the arousal is positive or negative. A person might be aroused because he is stressed and more likely to relapse, or because he is jubilant over a pleasant event such as watching his favorite team win a championship.

Another limitation of the current equipment and software is that the wireless transmission requires the individual to be within a few feet of a computer to download the physiologic data. This means that in most cases the equipment lacks the necessary mobility to detect arousal proactively when a person in recovery is living out his everyday life, navigating all of the life contexts in which stressful episodes could lead to relapse.

To overcome this, it would be necessary to transmit data wirelessly to a central server or another proximate device to detect changes in physiologic arousal on a close to real-time basis. This capability would provide the flexibility to follow people wherever they go and trigger needed supports whenever they might be helpful.

“All of the individual technology components for creating a product to decrease the likelihood of relapse based on biomonitoring data are available today,” says David Gustafson, PhD, professor of industrial engineering at the University of Wisconsin and principal investigator of the Innovations for Recovery initiative. “However, nobody that we are aware of has affordably integrated these various components so we can proactively use the data to provide the support that people in recovery need to stay abstinent and avoid relapse.” Gustafson and colleagues are seeking funding and partners to help them test these technologies to achieve these goals.

Bret R. Shaw, PhD, is an Assistant Professor in Life Sciences Communication at the University of Wisconsin-Madison, where he has been researching technology-based solutions to improve outcomes in the addiction treatment field with a focus on relapse prevention. His e-mail address is Bret_Shaw@chsra.wisc.edu.
Andrew Isham is a Research Assistant at the Center for Health Enhancement Systems Studies and a PhD student in Industrial and Systems Engineering at the University of Wisconsin.

References

1. Donovan DM. Assessment issues and domains in the prediction of relapse. Addiction 1996 Dec Suppl; 91:S29–36.
2. Witkiewitz K, Marlatt GA. Relapse prevention for alcohol and drug problems: that was Zen, this is Tao. Am Psychol 2004 May-Jun; 59:224–35.
3. Dennis ML, Scott CK, Funk R. An experimental evaluation of recovery management check-ups (RMC) for people with chronic substance abuse disorders. Eval Progr Plan 2003; 26:339–52.
4. McKay JR, Weiss RV. A review of temporal effects and outcome predictors in substance abuse treatment studies with long-term follow-ups: preliminary results and methodological issues. Eval Rev 2001 Apr; 25:113–61.
5. McLellan AT, Lewis DC, O'Brien CP, et al. Drug dependence, a chronic mental illness: implications for treatment, insurance, and outcomes evaluation. JAMA 2000; 284:1689–95.
6. McKay JR, Franklin TR, Patapis N, et al. Conceptual, methodological, and analytical issues in the study of relapse. Clin Psychol Rev 2006; 26:109–27.
7. Zywiak WJ, Westerberg VS, Conners GJ, et al. Exploratory findings from the Reasons for Drinking Questionnaire. J Subst Abuse Treat 2003 Dec; 25:287–92.
8. Baker TB, Piper ME, McCarthy DE, et al. Addiction motivation reformulated: an affective processing model of negative reinforcement. Psychol Rev 2004; 111:33–51.
9. Shaw B. Can new technologies prevent relapse?. Behav Healthc 2006; 26 (6): 30–1.