Can Warning Signs Training help maintain abstinence?

Alcohol Insight Number 30


A minority of alcohol dependent people resume drinking again and again, despite receiving help that is generally effective and having a strong desire not to drink. A different approach, designed specifically to help them, is widely used in the USA but not in the UK, and has never been properly evaluated. This approach suggests that many who resume drinking repeatedly exhibit an idiosyncratic set of changes, of which they are usually unaware, during the days before they start drinking again. The intervention aims to help people to discover their own habit­ual ‘warning signs’, and act on them to abort the unravelling process. There are published treatment manuals, DVDs and supporting material available that UK services wishing to offer this intervention can utilise.

They tested the claim that this novel intervention would help people who were dependent on alcohol and who had stopped drinking and then resumed drinking a number of times in the past. Success was defined as not drinking at all over the course of a year.

Their randomised trial studied people with a history of repeated re­lapse who had just completed day treatment which aimed to help them abstain from drinking. Half received the unit’s Routine Aftercare and half received this plus the Warning Signs intervention. Routine After­care encouraged participants to attend specific aftercare groups and social activities for as long as they wished.

Researchers followed up 93% of participants 12 months after they entered the trial and 85% after 24 months. The research hypothesis was that those receiving Warning Signs treatment would be less likely to resume drinking during follow-up.


  • During the first follow-up year 45% of the Warning Signs group and 26% of Routine Aftercare never drank heavily at all (‘heavy drinking’ being three consecutive days of drinking nine or more units of alco­hol a day). This difference was statistically significant. The Warning Signs intervention halved participants’ odds of drinking heavily.
  • For every five participants receiving the Warning Signs intervention one who would otherwise have drank heavily did not do so.
Fig 1

Never Drank Heavily During Follow-up Year

  • During this year 31% of the Warning Signs group and 17% of the Rou­tine Aftercare group never drank at all. The Warning Signs interven­tion halved participants’ odds of drinking during this time (although the difference was not statistically significant).

Never Drank At All During the Following Year

  • For every seven participants who entered Warning Signs one who would otherwise have drank did not do so. The Warning Signs group drank less often and also drank heavily less often during this year: these differences were significant.
  • The Warning Signs intervention was acceptable to service users: 96% of those eligible to join the trial did so, and those allocated to this condition attended a median of 13.5 of 15 individual sessions.
  • The differences in drinking between the two groups had disappeared by the second year follow-up, when their outcomes were similar.
  • Some 42% of all participants never drank during the 4 months prior to the two-year follow-up; 70% of all participants never drank heav­ily during those four months.
  • It would have been difficult to detect a treatment effect at the 24- month follow-up because significant numbers of those who resumed drinking stopped again during the second year, This dynamic picture could obscure any treatment effect that occurred.
  • Attending aftercare during the first month after completing day treatment was associated with better drinking outcomes during the following 11 months. Each aftercare session attended was associ­ated with 10 extra days of abstinence during the first year. These observations suggest, but cannot prove, that attending aftercare im­mediately after completing treatment caused a better outcome and should therefore be promoted.


  • The Warning Signs intervention caused reductions in heavy drinking sufficient to produce benefits to health.
  • These results justify using this intervention in similar settings to reduce the rate of resuming heavy drinking.
  • The effects are worthwhile because the participants were chosen for their history of resuming drinking after treatment and because the control group received a great deal of aftercare support and help.
  • The Warning Signs intervention achieved an effect in addition to the aftercare support. It would be valuable to replicate this single evaluation of the Warning Signs intervention in a very different set­ting, where there was less aftercare support.

Research Team

Dr Gerald Bennett, Jackie Withers, David Higgins, James Bailey, and Lorraine Parry of Dorset HealthCare NHS Trust in Bournemouth carried out the first evaluation of this intervention in collaboration with Profes­sor Peter Thomas of Bournemouth University.


This Alcohol Insight
The Final Report