Intervention uses smart phones to predict relapse in alcohol dependence

Mobile technology can be used to predict which patients with alcohol dependence are more likely to relapse and possibly prevent such relapses from occurring, results of a preliminary study of 152 adults suggest.

"This prediction algorithm may help addiction counselors take a proactive approach with these patients," wrote Ming-Yuan Chih, a doctoral candidate at the University of Wisconsin, Madison, and his colleagues. "Information and mobile communication technology offer a great opportunity to develop novel ways to deliver addiction treatment to patients."

The investigators recruited and randomized 170 patients from two residential treatment organizations – one in the northeastern United States, the other in the Midwest. The mean age of the patients was 38 years; most were white and male. Patients with either a significant developmental or cognitive impairment or a history of suicidality were excluded. More than half of the patients had abused drugs beyond alcohol in their lives, and almost half reported mental health problems (J. Subst. Abuse Treat. 2014;46:29-35).

© Nick Smith/iStockphoto

Before leaving treatment, the participants in the intervention group were given smart phones with a mobile broadband connection and training on how to use the addiction-comprehensive health enhancement support, or A-CHESS system. They were expected to submit a "Weekly Check-In" to A-CHESS once every 7 days, and those numbers were sent to a secure server at the university. Seven days after each submission, the patients received a prompt to submit another Weekly Check-In. To remain in the trial, however, Weekly Check-Ins were not required. "Therefore, it is possible that substantially more than 7 days could elapse between submissions," the investigators wrote. In the end, 152 of the patients submitted 2,934 Weekly Check-In reports between April 2010 and August 2011.

The check-in survey consisted of three screens. The first asked the patients whether they had abused drugs or alcohol in the last 7 days. If the answer was yes, the patients were asked whether they wanted their A-CHESS counselor to be notified. The second screen sought to pinpoint other experiences that patients might have had in the last week. For example, one question asked whether they had had difficulty sleeping on a scale of 0-7. Another asked the patients to rate their "level of depression," and yet another asked them to rate their "drinking urges." The last screen sought to determine how patients had been spending their time over the last 7 days. They were asked to rate their Alcoholics Anonymous meeting attendance and their involvement with spiritual activities and with work and school. The intervention lasted for 8 months.

The investigators used data from screens #2 and #3 to determine each patient’s recovery progress score. This score could range from –35 to +35, and the higher scores were tied to better outcomes. In addition, the investigators used the data to construct each patient’s lapse history. Using these data, the investigators developed a model that enabled them to determine the chances of a patient lapsing within 1 week. The model was used in A-CHESS to "identify patients at high risk and then to take tailored action to reduce that risk," Mr. Chih and his colleagues wrote.

Patients with recovery progress scores of –35 to –28 had a 33% probability of lapse and a 67% probability of non-lapse. Those with recovery progress scores of +28 to +35 had a 50/50 probability of lapse and non-lapse. Those who had a good chance of relapsing within the following week automatically received a text message about the risk. The text included suggestions about steps the patient could take to avoid relapse within A-CHESS, such as planning alternative activities. The patient’s counselor also received a text message about the possible relapse and could reach out to the patient at that point. Finally, the A-CHESS study coordinator would receive a text alert. Mr. Chih noted that this model – including the alerting feature – has been implemented as part of A-CHESS.

Several limitations were cited. Because only two treatment organizations were used in this study, it is unclear whether the model is generalizable. Also, each patient’s lapse status was self-reported and might not be accurate. Still, these results serve as a "starting point for further development," the researchers wrote. "We need to find better ways to assist people with the very difficult transition from alcohol addiction to sobriety."

The National Institute on Alcohol Abuse and Alcoholism funded the study. Mr. Chih reported no conflicts of interest.

[email protected]

Mobile technology can be used to predict which patients with alcohol dependence are more likely to relapse and possibly prevent such relapses from occurring, results of a preliminary study of 152 adults suggest.

"This prediction algorithm may help addiction counselors take a proactive approach with these patients," wrote Ming-Yuan Chih, a doctoral candidate at the University of Wisconsin, Madison, and his colleagues. "Information and mobile communication technology offer a great opportunity to develop novel ways to deliver addiction treatment to patients."

The investigators recruited and randomized 170 patients from two residential treatment organizations – one in the northeastern United States, the other in the Midwest. The mean age of the patients was 38 years; most were white and male. Patients with either a significant developmental or cognitive impairment or a history of suicidality were excluded. More than half of the patients had abused drugs beyond alcohol in their lives, and almost half reported mental health problems (J. Subst. Abuse Treat. 2014;46:29-35).

© Nick Smith/iStockphoto

Before leaving treatment, the participants in the intervention group were given smart phones with a mobile broadband connection and training on how to use the addiction-comprehensive health enhancement support, or A-CHESS system. They were expected to submit a "Weekly Check-In" to A-CHESS once every 7 days, and those numbers were sent to a secure server at the university. Seven days after each submission, the patients received a prompt to submit another Weekly Check-In. To remain in the trial, however, Weekly Check-Ins were not required. "Therefore, it is possible that substantially more than 7 days could elapse between submissions," the investigators wrote. In the end, 152 of the patients submitted 2,934 Weekly Check-In reports between April 2010 and August 2011.

The check-in survey consisted of three screens. The first asked the patients whether they had abused drugs or alcohol in the last 7 days. If the answer was yes, the patients were asked whether they wanted their A-CHESS counselor to be notified. The second screen sought to pinpoint other experiences that patients might have had in the last week. For example, one question asked whether they had had difficulty sleeping on a scale of 0-7. Another asked the patients to rate their "level of depression," and yet another asked them to rate their "drinking urges." The last screen sought to determine how patients had been spending their time over the last 7 days. They were asked to rate their Alcoholics Anonymous meeting attendance and their involvement with spiritual activities and with work and school. The intervention lasted for 8 months.

The investigators used data from screens #2 and #3 to determine each patient’s recovery progress score. This score could range from –35 to +35, and the higher scores were tied to better outcomes. In addition, the investigators used the data to construct each patient’s lapse history. Using these data, the investigators developed a model that enabled them to determine the chances of a patient lapsing within 1 week. The model was used in A-CHESS to "identify patients at high risk and then to take tailored action to reduce that risk," Mr. Chih and his colleagues wrote.

Patients with recovery progress scores of –35 to –28 had a 33% probability of lapse and a 67% probability of non-lapse. Those with recovery progress scores of +28 to +35 had a 50/50 probability of lapse and non-lapse. Those who had a good chance of relapsing within the following week automatically received a text message about the risk. The text included suggestions about steps the patient could take to avoid relapse within A-CHESS, such as planning alternative activities. The patient’s counselor also received a text message about the possible relapse and could reach out to the patient at that point. Finally, the A-CHESS study coordinator would receive a text alert. Mr. Chih noted that this model – including the alerting feature – has been implemented as part of A-CHESS.

Several limitations were cited. Because only two treatment organizations were used in this study, it is unclear whether the model is generalizable. Also, each patient’s lapse status was self-reported and might not be accurate. Still, these results serve as a "starting point for further development," the researchers wrote. "We need to find better ways to assist people with the very difficult transition from alcohol addiction to sobriety."

The National Institute on Alcohol Abuse and Alcoholism funded the study. Mr. Chih reported no conflicts of interest.

[email protected]

Mobile technology can be used to predict which patients with alcohol dependence are more likely to relapse and possibly prevent such relapses from occurring, results of a preliminary study of 152 adults suggest.

"This prediction algorithm may help addiction counselors take a proactive approach with these patients," wrote Ming-Yuan Chih, a doctoral candidate at the University of Wisconsin, Madison, and his colleagues. "Information and mobile communication technology offer a great opportunity to develop novel ways to deliver addiction treatment to patients."

The investigators recruited and randomized 170 patients from two residential treatment organizations – one in the northeastern United States, the other in the Midwest. The mean age of the patients was 38 years; most were white and male. Patients with either a significant developmental or cognitive impairment or a history of suicidality were excluded. More than half of the patients had abused drugs beyond alcohol in their lives, and almost half reported mental health problems (J. Subst. Abuse Treat. 2014;46:29-35).

© Nick Smith/iStockphoto

Before leaving treatment, the participants in the intervention group were given smart phones with a mobile broadband connection and training on how to use the addiction-comprehensive health enhancement support, or A-CHESS system. They were expected to submit a "Weekly Check-In" to A-CHESS once every 7 days, and those numbers were sent to a secure server at the university. Seven days after each submission, the patients received a prompt to submit another Weekly Check-In. To remain in the trial, however, Weekly Check-Ins were not required. "Therefore, it is possible that substantially more than 7 days could elapse between submissions," the investigators wrote. In the end, 152 of the patients submitted 2,934 Weekly Check-In reports between April 2010 and August 2011.

The check-in survey consisted of three screens. The first asked the patients whether they had abused drugs or alcohol in the last 7 days. If the answer was yes, the patients were asked whether they wanted their A-CHESS counselor to be notified. The second screen sought to pinpoint other experiences that patients might have had in the last week. For example, one question asked whether they had had difficulty sleeping on a scale of 0-7. Another asked the patients to rate their "level of depression," and yet another asked them to rate their "drinking urges." The last screen sought to determine how patients had been spending their time over the last 7 days. They were asked to rate their Alcoholics Anonymous meeting attendance and their involvement with spiritual activities and with work and school. The intervention lasted for 8 months.

The investigators used data from screens #2 and #3 to determine each patient’s recovery progress score. This score could range from –35 to +35, and the higher scores were tied to better outcomes. In addition, the investigators used the data to construct each patient’s lapse history. Using these data, the investigators developed a model that enabled them to determine the chances of a patient lapsing within 1 week. The model was used in A-CHESS to "identify patients at high risk and then to take tailored action to reduce that risk," Mr. Chih and his colleagues wrote.

Patients with recovery progress scores of –35 to –28 had a 33% probability of lapse and a 67% probability of non-lapse. Those with recovery progress scores of +28 to +35 had a 50/50 probability of lapse and non-lapse. Those who had a good chance of relapsing within the following week automatically received a text message about the risk. The text included suggestions about steps the patient could take to avoid relapse within A-CHESS, such as planning alternative activities. The patient’s counselor also received a text message about the possible relapse and could reach out to the patient at that point. Finally, the A-CHESS study coordinator would receive a text alert. Mr. Chih noted that this model – including the alerting feature – has been implemented as part of A-CHESS.

Several limitations were cited. Because only two treatment organizations were used in this study, it is unclear whether the model is generalizable. Also, each patient’s lapse status was self-reported and might not be accurate. Still, these results serve as a "starting point for further development," the researchers wrote. "We need to find better ways to assist people with the very difficult transition from alcohol addiction to sobriety."

The National Institute on Alcohol Abuse and Alcoholism funded the study. Mr. Chih reported no conflicts of interest.

[email protected]