A client finishes twelve sessions of cognitive behavioral therapy for alcohol use. Drinking has stopped, mood has lifted, and both client and clinician feel the work is done. Six months later, the client returns. Or the client never returns at all, convinced that treatment simply failed.

This pattern is common across presenting problems. Addiction, depression, anxiety disorders, and eating disorders all show substantial relapse rates after successful acute treatment. The cause is rarely that the original intervention was ineffective. More often, treatment was designed to produce change and quietly assumed that change would sustain itself.

Relapse prevention, developed by G. Alan Marlatt and Judith Gordon in the 1980s, challenged that assumption. It treats maintenance as a distinct phase, with its own psychological mechanisms and its own skill set. Three elements of the model are especially useful in practice: understanding how a single lapse can escalate, mapping high-risk situations before they arrive, and building a life that makes a return to old patterns less attractive.

Marlatt and Gordon drew a distinction that sounds simple but carries significant clinical weight. A lapse is a single slip. A relapse is a return to the previous pattern. The question that matters is what happens between the two. Their answer was the abstinence violation effect, a cognitive and emotional reaction that can turn one drink, one binge, or one avoided situation into a full return to baseline.

The effect has two components. The first concerns attribution. When a client explains a lapse through internal, stable, and global causes, the slip becomes evidence of a fixed personal deficit. Typical examples are "I have no willpower" or "I never really changed." The second component is emotional. The gap between a self-image as someone in recovery and the behavior just enacted produces guilt and shame. Paradoxically, the old behavior is often the most familiar way to relieve those feelings.

Consider a client with a history of binge eating who eats more than planned at a family gathering. If she concludes that "the plan is ruined anyway," continued eating follows a certain logic. All-or-nothing rules often feel motivating during initial change, yet they create exactly the conditions for this cascade. The stricter the rule, the more total its violation appears.

Clinically, the intervention begins before any lapse occurs. Clinicians can present lapses as probable events and frame them as information rather than verdicts. They can also rehearse alternative attributions that are external, specific, and changeable, such as "I was exhausted and unprepared for that situation." Some practitioners develop a written lapse plan with the client. This is a short card describing what to do in the first hours after a slip, including whom to contact and what to tell oneself.

Takeaway

A lapse becomes a relapse largely through the story told about it. Preparing that story in advance is one of the most protective things treatment can do.

The abstinence violation effect explains what happens after a slip. High-risk situations explain why slips occur in the first place. Marlatt's early research found that most relapse episodes clustered around a few situational categories: negative emotional states, interpersonal conflict, and social pressure. These categories are useful starting points. Effective maintenance planning, however, requires a map that is specific to the individual.

The core assessment tool is functional analysis. Clinician and client review past lapses in detail. They establish where the client was, who was present, what the client felt, what thoughts preceded the behavior, and what consequences followed. Self-monitoring logs kept during treatment add data as new situations arise. Structured instruments such as the Inventory of Drinking Situations can supplement this process. Even so, the clinical interview usually surfaces the most personally meaningful triggers.

Marlatt also identified a subtler pathway he called apparently irrelevant decisions. These are small, seemingly reasonable choices that gradually move someone toward a high-risk situation. A client recovering from depression might skip one morning walk because of rain, then cancel coffee with a friend, then spend a weekend alone. No single decision looks like relapse. Together, they recreate the conditions under which symptoms first took hold. Teaching clients to recognize these early links in the chain gives them more time to intervene.

Once the map exists, client and clinician develop a coping plan for each situation and rehearse it until the client can reach for it under stress. This matters because high-risk moments are, by definition, the times when planning capacity is lowest. A plan created calmly in session and practiced through role-play or imagery is far more likely to be used than one improvised in the moment. Later work by Witkiewitz and Marlatt emphasized that these risk factors interact dynamically, so maps should be revisited as circumstances change.

Takeaway

Coping skills are least available precisely when they are most needed. The work of maintenance is to decide in calm moments what you will do in difficult ones.

Situational planning addresses acute risk. Marlatt argued that relapse also has a slower, background cause, which he described as an imbalance between "shoulds" and "wants." Shoulds are obligations experienced as external demands. Wants are activities that bring genuine pleasure or meaning. When obligations dominate, people build up a sense of deprivation and entitlement that makes the old behavior feel deserved. A common version is "After the week I've had, I've earned this."

This insight shifts the focus of maintenance from preventing a behavior to building a life. Removing a problem behavior often leaves a gap. Alcohol may have filled evenings, structured social life, or regulated stress. Avoidance in anxiety may have protected time and energy. If nothing replaces those functions, the pull back toward the familiar remains constant, even when no specific trigger is present.

Assessment here looks at the overall shape of a client's week. Activity scheduling, borrowed from behavioral activation for depression, shows concretely how time is divided between obligation, rest, pleasure, and meaningful engagement. Values clarification, drawn from approaches such as acceptance and commitment therapy, helps identify activities that might sustain motivation over years rather than weeks. Marlatt encouraged what he called positive addictions. These are regular practices, such as exercise or meditation, that become rewarding in their own right and so keep themselves going.

Mindfulness-Based Relapse Prevention, developed by Bowen, Chawla, and Marlatt, combines this thinking with present-moment awareness. Clients learn to notice cravings and urges as passing experiences rather than commands. At the same time, they cultivate a life in which those urges have less to feed on. A similar logic underpins mindfulness-based cognitive therapy for recurrent depression. In both approaches, the aim is not constant vigilance. It is a broader foundation of wellbeing that makes vigilance less necessary.

Takeaway

Sustained change is rarely held in place by willpower alone. It is held in place by a life that offers something better than what was given up.

Relapse prevention reframes the end of treatment as a transition rather than a finish line. The skills that produce initial change differ from the skills that sustain it, so maintenance deserves explicit time in the treatment plan.

In practice, this means introducing relapse prevention well before termination. Normalize lapses and rehearse adaptive attributions. Analyze past episodes and build individualized coping plans. Review the balance of obligations and rewards in the client's life, and help fill the gaps that symptom reduction leaves behind. Booster sessions can consolidate these gains.

Perhaps the most useful shift is in how success is measured. A client who lapses, recovers quickly, and learns something from it has not failed treatment. That client has used it.