Can mobile apps help recovering alcoholics?

In a previous blog regarding smoking cessation I observed that quitting tobacco is easier than staying quit with 75% of quitters relapsing within 12 months (Agboola, 2009).

Unfortunately the same is true for alcohol use disorders. Indeed, relapse rates following detoxification can be as high as 80% if there is no continuing treatment (Moyer and Finney, 2002). This may be particularly problematic in regard to alcohol use as the medical consequences of detoxification can escalate with multiple withdrawals (Lechtenberg and Worner, 1990).

It is notable that relapse rates are reduced if there is some continuous pharmacological (Rösner et al, 2010) or cognitive (Carroll, 1996) treatment following inpatient detoxification. So obviously, continued contact seems to be the solution, but with the NHS already under financial strain how much time and resources can be devoted to such ongoing treatment?

One possible solution is to harness hand-held technology such as smartphones to deliver flexible, continuous support for recovering alcoholics. This, theoretically, could offer cost-effective continuing care, that will reduce relapse rates and ultimately improve the chances of continued abstinence.

Recently a team of US-based researchers conducted a randomised controlled trial (RCT) into the efficacy of a smartphone app: Addiction-Comprehensive Health Enhancement Support System (A-CHESS). They used a sample of alcohol dependent individuals who were about to leave residential treatment. The app was designed to offer flexible emotional and instrumental support if and when the patient needs it.


The A-CHESS app is designed to provide emotional and instrumental support to patients when they need it.


The current study was an RCT of 349 alcohol dependent patients who had been in a residential treatment program.

The five treatment centres involved all made use of cognitive behavioural therapy (CBT), motivational interviewing and psychoeducation (all conducted in groups) although two of the centres also included case management and individual counselling. Patients were randomised into the A-CHESS condition or treatment as normal, which did not include continuous aftercare.

The A-CHESS app offers a range of different features including a panic button when individuals feel tempted to drink alcohol, which gives them different options to deal with temptation or stress, a high risk location alert and audio relaxation guides. Examples of the services in the app can be found on the University of Wisconsin-Madison CHESS website).

The primary outcome was risky drinking days, defined as 4+ (males) or 3+ (females) standard drinks in an hour. The authors also investigated the impact of A-CHESS on abstinence and negative consequences of drinking (job loss, arrest etc). These outcomes were reported at 4, 8 and 12 months post-residential treatment.

Finally, perceived confidence in avoiding alcohol use (drug-taking confidence questionnaire), relatedness (important people survey) and autonomous motivation to control drinking (self-regulation questionnaire) were measured as possible mediators of the effect the app has on alcohol use.

Patients were randomized into the A-CHESS condition or treatment as normal, which did not include continuous aftercare.

Patients were randomised into the A-CHESS condition or treatment as normal, which did not include continuous aftercare.


Patients receiving the A-CHESS program had significantly fewer overall risky drinking days (1.39, SE= 0.34) than those in the control condition (2.75, SE= 0.34). This overall effect was driven by significant differences at 4 months and 12 months, although there was no group difference at 8 months. However it is important to note that although significant effect sizes were all small for all these group differences.

Relapse was lower in the A-CHESS group at months 8 (OR 1.70: 95% CI 1.05 to 2.96), 12 (OR 1.94: 95% CI 1.14 to 3.31) and overall (OR 1.65: 95% CI 1.05 to 2.57), although there was no difference at 4 months. There was no effect on negative consequences of drinking.

Finally, the mediation analysis found that only changes in perceived confidence (i.e. scores on drug-taking confidence questionnaire) mediated the effect of A-CHESS on risky drinking days, with this effect at four months only.


Overall the findings do seem promising, with statistically significant effects on risky drinking and relapse at multiple time points. Indeed, using smartphones to deliver treatment when it is needed most seems like a cost effective intervention that may reduce the burden on the NHS.


However, there are some caveats to this conclusion:

  1. Some refinement is needed, effect sizes for the intervention are small so any gains are actually fairly limited.
  2. It is important to remember that although we view smartphones as ubiquitous there is still a digital divide in the UK, suggesting that such interventions will be more useful to some groups than others. Indeed many people may have no familiarity with smart phones and struggle with such technology.
  3. I would recommend treating the mediation analysis very cautiously; all mediators are considered separately and at each time point (and possibly overall, the paper is not clear). This suggests the authors ran at least nine joint significance tests, with only one finding evidence for mediation. Furthermore, the authors do not report confidence intervals for the significant effect described.  I would be reticent to make any firm conclusions based upon such an analysis.
Can mobile apps help recovering alcoholics? Image: Ed Yourdon CC BY-NC-SA 2.0

Can mobile apps help recovering alcoholics?
Image: Ed Yourdon CC BY-NC-SA 2.0


Gustafson DH, McTavish FM, Chih M, Atwood AK, Johnson RA, Boyle MG, Levy, SM, Driscoll H, Chisholm, SM, Dillenburg L, Isham A, and Shah D (2014) A Smartphone Application to Support Recovery From Alcoholism: A Randomized Clinical TrialJAMA Psychiatry. doi:10.1001/jamapsychiatry.2013.4642

Agboola SA, Coleman TJ and  McNeill AD (2009) Relapse prevention in UK Stop Smoking Services: a qualitative study of health professionals’ views and beliefs. BMC Health Services Research 9:67 doi:10.1186/1472-6963-9-6

Moyer A and Finney JW (2002) Outcomes for untreated individuals involved in randomized trials of alcohol treatment. Journal of Substance Abuse Treatment 23: 2 DOI: 10.1016/S0740-5472(02)00264-7. [PubMed abstract]

Lechtenberg R and Worner TM (1990) Seizure Risk With Recurrent Alcohol Detoxification Archives of Neurology 47: 5. doi:10.1001/archneur.1990.00530050055012. [PubMed abstract]

Rösner S, Hackl-Herrwerth A, Leucht S, Lehert P, Vecchi S, Soyka M (2010). Acamprosate for alcohol dependence. Cochrane Database of Systematic Reviews 9. Art. No.: CD004332. DOI: 10.1002/14651858.CD004332.pub2.

Carroll, K M. (1996) Relapse prevention as a psychosocial treatment: A review of controlled clinical trials. Experimental and Clinical Psychopharmacology, 4:1, doi: 10.1037/1064-1297.4.1.46 [Abstract]

Share on Facebook Tweet this on Twitter Share on LinkedIn Share on Google+