Contingency management for cocaine: the evidence: Defined target; Verified progress; Positive reinforcement.
Article overview by AddictionResearch. Read the findings, context and limitations below.
Evidence at a glance
Design
Systematic review and meta-analysis; clinical guideline
Period / population
Meta-analysis 2021; ASAM/AAAP guideline 2024
Review status
AI-assisted educational summary; no independent clinician review.
The key finding

Contingency management has an important evidence base in stimulant treatment. Ask how a service actually delivers it, measures progress and connects it with continuing care.

What the meta-analysis found

A 2021 meta-analysis by Bentzley and colleagues included 157 studies and 15,842 participants. Contingency management was the only treatment category consistently associated with increased odds of a cocaine-negative urine result across the main and sensitivity analyses (odds ratio 2.13; 95% confidence interval 1.62–2.80).

The analysis compared end-of-treatment results with baseline across heterogeneous studies. It should not be read as a single head-to-head trial showing that a particular clinic doubles recovery. Its measured outcome was also narrower than long-term health, functioning or sustained recovery.

What contingency management means

Contingency management uses a defined positive reinforcement process linked to a specific behaviour, such as verified abstinence or attendance. The 2024 ASAM/AAAP stimulant guideline identifies it as a primary component of treatment, alongside other psychosocial approaches.

A clear programme explains the target, verification process, timing and reinforcement rules. It also explains how people are supported when a target is not met. The intervention is a structured treatment approach; a vague promise of rewards is not enough to establish that the studied model is being delivered.

Read the endpoint before the headline

A negative sample at one time point does not describe every day before or after it. Attendance, treatment completion and biological testing each answer different questions. A useful outcomes report shows how missed tests are handled and whether results include all participants or only those retained.

Consider a hypothetical study with frequent testing in one group and little testing in another. Raw counts of negative samples would be difficult to compare without knowing the opportunities to provide a sample. Always look for a definition that makes the comparison meaningful.

Questions for a prospective programme

  • Is contingency management available for my treatment needs, and what behaviour does it target?
  • How often are progress and testing reviewed?
  • What happens after a missed visit or a positive result?
  • Which psychological and practical supports accompany the programme?
  • How will care continue when the structured intervention ends?

Ask staff to describe a normal week and an example of how they respond when progress is uneven. That answer is usually more informative than a brochure listing treatment names.

How much can we conclude?

The combined evidence supports taking the intervention seriously. It does not establish that it is sufficient on its own for everyone, or remove the need to assess mental health, other substances, physical symptoms and the person’s circumstances. Availability and delivery can differ by service and jurisdiction.

For preparation and safety concerns, read the cocaine guide. For broader claims about programme outcomes, see how to interpret rehab success rates.

Sources and further reading

Sources checked on 18 September 2026. Publication and data years are identified separately.

  1. Bentzley et al. Comparison of treatments for cocaine use disorder among adults. JAMA Network Open (2021). DOI: 10.1001/jamanetworkopen.2021.8049.
  2. ASAM/AAAP. Clinical practice guideline on the management of stimulant use disorder (2024, PDF).
About this page

Published by AddictionResearch with AI-assisted preparation. General education; no independent clinician review is claimed.

Editorial standards Report a correction