Drug and Alcohol Dependence 216 (2020) 108307

Contents lists available at ScienceDirect

Drug and Alcohol Dependence

journal homepage: www.elsevier.com/locate/drugalcdep

Contingency management for the treatment of methamphetamine use

disorder: A systematic review
Hayley D. Brown, Anthony DeFulio *
Western Michigan University, United States

A R T I C L E I N F O A B S T R A C T

Keywords: Methamphetamine use continues to be an important public health problem. Contingency management is among
Methamphetamine the most effective interventions for reducing methamphetamine use. It has been more than ten years since the
Contingency management last systematic review of contingency management for methamphetamine use disorder. Since then, an additional
Outpatient treatment
ten randomized controlled trials and a variety of other studies have been completed. The present systematic
Predictors
review includes 27 studies. Several factors, most notably problem severity, appear to predict treatment outcome.
However, the effectiveness of CM has been demonstrated in studies restricted to MSM, studies restricted to
implementation in community programs, and in studies of the general population of methamphetamine users
conducted in research treatment programs. There appear to be broad benefits of contingency management
intervention, including greater drug abstinence, higher utilization of other treatments and medical services, and
reductions in risky sexual behavior. Twenty of the twenty-one studies that reported abstinence outcomes showed
an effect of contingency management on abstinence, and seven of the nine studies that reported sexual risk
behavior outcomes showed an effect of contingency management in reducing risky sexual behavior. Taken
together, recent evidence suggests strongly that outpatient programs that offer treatment for methamphetamine
use disorder should prioritize adoption and implementation of contingency management intervention.

1. Introduction reduces stimulant use in general and cocaine use in particular (e.g.,
Higgins et al., 1991, 1993, 2000). However, cocaine and methamphet­
Methamphetamine use is an important public health problem in the amine have different drug effects and consequences. Relative to cocaine
United States that raises serious medical and criminal justice concerns. users, methamphetamine users are more likely to be people of color, and
In 2018 approximately 1.9 million Americans reported using metham­ more likely to experience severe medical and psychiatric consequences
phetamine (Substance Abuse and Mental Health Services Administra­ (Rawson et al., 2000). Nevertheless, prior reviews of psychosocial
tion, 2019), and in the age-adjusted rate of overdose deaths involving intervention for methamphetamine use disorder suggest that contin­
psychostimulants with abuse potential, including methamphetamine, gency management intervention outcomes for methamphetamine users
had increased nearly five times its 2012 level (Hedegaard et al., 2020). parallel those of cocaine users (Baker and Lee, 2003; Lee and Rawson,
Currently, there are no widely-approved pharmacotherapies to 2008). The review by Baker and Lee recommended incorporation of
curtail methamphetamine use. However, behavioral interventions have contingency management programs into residential and outpatient
proved promising (Rawson et al., 2006; Roll et al., 2006b). The most treatments for methamphetamine use based on the findings of a single
effective intervention to date appears to be contingency management, a study that was conducted in an inpatient setting (Melin and Götestam,
behavioral intervention in which material incentives are delivered 1973). The review by Lee and Rawson included a section on contingency
contingent on biological confirmation of drug abstinence. Contingency management that featured five studies, and a section on
management is effective in promoting abstinence from a variety of cognitive-behavior therapy approaches for specific groups of users that
substances including alcohol, opioids, and stimulants (Prendergast et al., included three studies from a data set drawn from a controlled trial in
2006). which participants were assigned to various combinations of
Many studies have shown that contingency management effectively cognitive-behavior therapy and contingency management. Many of the

* Corresponding author at: 1903 W Michigan Ave, Kalamazoo MI, 49008-5439, United States.
E-mail address: [email protected] (A. DeFulio).

https://doi.org/10.1016/j.drugalcdep.2020.108307
Received 3 June 2020; Received in revised form 13 September 2020; Accepted 16 September 2020
Available online 21 September 2020
0376-8716/© 2020 Elsevier B.V. All rights reserved.
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

eight studies include in that review included analyses of data sets that that evidence from randomized controlled trials is presented first, and
were drawn from a broader population of stimulant users, and in some other studies are presented in a second section. Within these sections,
cases were predominantly cocaine users. Despite these limitations, the the studies are organized by the specific outcomes reported in the con­
authors concluded that contingency management studies produced tingency management interventions.
substantial decreases in methamphetamine use during treatment. Sub­
stantial progress has been made in the evaluation of contingency man­ 2. Methods
agement interventions for methamphetamine use since the time of those
reviews. Some of the more recent contingency management studies in 2.1. Search strategy
which methamphetamine use was targeted were included in a recent
systematic review of non-pharmacological interventions for metham­ MEDLINE, PsycINFO, and Web of Science databases were searched
phetamine use (AshaRani et al., 2020). This review included seven using the keywords “contingency management” AND “methamphet­
studies, only one of which was included in the Lee and Rawson review, amine.” There was no start date limitation on the search, and the last
and indicated that all seven showed a decrease in drug use, and that two search date for inclusion was conducted in August 2020. These articles
also showed improvements in other health behaviors. The discussion were evaluated for inclusion independently by both authors. Twenty-six
also mentions other effects of contingency management, including articles were found to meet the inclusion criteria. References within and
enhanced retention in treatment and improvements in psychological citations to all included articles were hand searched by the first author to
symptoms, though these findings are not described in detail. identify additional articles. This process yielded one more study for in­
The prior reviews of non-pharmacological interventions for meth­ clusion. That article did not yield additional articles, resulting 27
amphetamine use are helpful and clearly illustrate the efficacy of con­ included studies. This process is outlined in a flowchart (see Fig. 1) and a
tingency management in reducing methamphetamine use. The literature detailed listings of the included studies and their characteristics are
as a whole now contains important information about the effects of presented in Tables 1 and 2.
contingency management on drug abstinence, drug craving, treatment
engagement and retention, earning and spending of incentives, mood/
affect, sexual risk behaviors, and intervention acceptability. The goal of 2.2. Eligibility criteria
the present review is to provide a fuller, more complete and detailed
review that is focused specifically on contingency management as an Studies were included if they contained a contingency management
intervention for methamphetamine use. The review is organized such intervention that targeted methamphetamine abstinence and the ma­
jority of participants were methamphetamine users. All contingency

Fig. 1. Flowchart of the article search process.

2
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

Table 1
Summary of Randomized Controlled Trials.
Author (year) N Inclusion/Exclusion Criteria Intervention Description Outcomes Relevant Findings

Carrico et al. 21 Inclusion: identify as male; Positive affect + CM or CM Positive and negative affect, Across groups, participants provided an
(2015) report having anal sex with a only for 12-weeks Follow-up: 3 self-reported stimulant use, total average of 30.5 (SD = 6.3) meth
man in the past year; and report and 6-month Design: RCT non-reactive urine toxicology negative samples out of 36 total
using meth at least weekly on results for stimulants during CM samples and provided an average of
average during the 3 months 20.7 (SD = 10.7) consecutive negative
prior to CM. samples. Positive affect + CM
participants reported an increase in
positive affect (B = 4.67, p < .05;
Cohen’s d = 0.24) at the end of
treatment. CM-only participants
reported a reduction in negative affect
(B=-7.69, p < .05; Cohen’s d=-0.23) at
the end of treatment.
Carrico et al. 110 Inclusion: 18 years of age or Positive affect + CM or Positive and negative affect, Greater reductions in self-reported
(2018) older, report of anal sex with a attention-control + CM for 12 mindfulness, methamphetamine stimulant use (Cohen’s d = − .46,
man in the past 12 months, weeks Follow-up: 3-month craving, self-reported stimulant p = .030) and methamphetamine
speak English, provide Design: RCT use, total non-reactive urine craving (Cohen’s d = − .51, p = .026)
documentation of HIV-positive toxicology results for stimulants were observed at 3 months for the
serostatus, provide a urine or during CM positive affect condition.
hair sample that was reactive
for meth
Chudzynski et al. 120 Inclusion: 18–65 years of age, Continuous reinforcement, abstinence verified by urine Compared to standard treatment
(2015) met DSM-IV criteria for meth intermittent predictable drug analysis results participants, continuous CM
dependence, were willing and reinforcement, intermittent participants were almost two times
able to comply with study unpredictable reinforcement, more likely to submit a negative meth
procedures, were willing and or standard condition for 16 UA (OR = 1.98, p < .05), intermittent
able to provide written weeks Follow-up: 10 and 12- predictable CM participants were 2.4
informed consent. Exclusion: weeks Design: RCT times more likely to submit a negative
had medical or psychiatric meth UA (OR = 2.40, p < .05), and
conditions that precluded safe intermittent unpredictable CM
study involvement, were unable participants were about 1.7 times more
to comply with the research likely to submit a negative meth UA
requirements, a history of (OR = 1.72, p < .05). None of the CM
violent criminal behavior or be conditions’ OR differed significantly
on parole, any other (p < .05) from one another.
circumstances that would
interfere with safe study
participation.
Corsi et al. 58 Inclusion: Meth use in the CM or CM + strengths-based Attendance at sessions, urine A significant reduction in amphetamine
(2012) previous 30 days (verified case management for 17 weeks results, total vouchers earned, and meth use was observed from
through observed urinalysis Follow-up: 4 an 8-month satisfaction with the baseline to each follow-up (χ2 = 11.6,
(UA) and self-report of meth Design: RCT intervention and reported p < 0.001). 100 % of samples at
use, reported sex with an barriers baseline were positive for meth, 53.3 %
opposite-sex partner in last 30 were positive at 4 and 8-month follow-
days, negative urinalysis for ups. Participants rated the intervention
opiates and methadone during highly and reported few barriers to
baseline screening period, session attendance.
ability to provide reliable
contact information, not in drug
treatment in the 30 days prior
to the baseline interview,
willing to be tested for HIV at
baseline and follow-up, not
transient and committed to
being available for follow-up
interviews.
Corsi et al. 253 Inclusion: 18 years of age or CM or CM + strengths-based Attendance at sessions, urine CM/SBCM was significantly associated
(2019) older and competent to give case management (CM/SBCM) results with attending at least one UA session
informed consent at the time of for 17 weeks Follow-up: 4 and (p-value = 0.0139). CM/SBCM
the interview, meth use 8-month Design: RCT participants were 2.7 times more likely
(verified through urine drug to attend at least one UA session than
screening and a self-report of partnered control group participants
meth use of at least 4 times per (95% CI 1.2, 6.0, p-value = 0.0014).
month for the last 3 months), Participants who earned more money
self-reported sex with someone during session A (weeks 1− 4) were
of the opposite sex in last 30 more likely to submit negative UAs
days, ability to provide a during session B (weeks 5− 17) (p-value
reliable address and phone <0.0001).
number for contact, not in drug
treatment in the past 30 days,
willingness to be tested for HIV
at baseline and follow-up, not
transient and no known reason
why he/she would not be
(continued on next page)

3
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

Table 1 (continued )
Author (year) N Inclusion/Exclusion Criteria Intervention Description Outcomes Relevant Findings

available for follow-up

interviews, not currently
mandated by the criminal
justice system to receive
treatment.
Menza et al. 127 Inclusion: age ≥18 years, CM or referral to community Report of unprotected anal CM and control participants were
(2010) willingness to be randomized resources for 12-weeks Follow- intercourse (UAI) with a partner comparably likely to provide urine
and provide locator up: 3-month Design: RCT of unknown or discordant HIV samples containing meth (adjusted
information, no plans to move status (non-concordant UAI) in relative risk [aRR] = 1.09; 95 %CI:
from the study catchment area the prior six weeks, number of 0.71, 1.56) and to report non-
within 6 months of enrollment, non-concordant UAI partners, concordant UAI (aRR = 0.80; 95%CI:
≥1 episodes of anal sex, ≥2 results of methamphetamine 0.47, 1.35). At follow-up: CM
episodes of meth use in the urine testing, self-reported participants were somewhat more
month prior to screening weekly or more frequent use of likely to provide urine samples
Exclusion: reported a mutually methamphetamine, self- containing meth than control
monogamous relationship with reported use of >8 quarters (two participants (aRR = 1.21; 95%CI: 0.95,
a partner of the same HIV status grams) of methamphetamine. 1.54, P = 0.11). CM participants were
lasting ≥2 years, men who more likely to report at least weekly
expressly asked for meth use and use of more than eight
detoxification, counseling, or quarters over the prior six weeks at the
drug rehabilitation services. time of measurement than control
participants during the intervention
and post-intervention periods.
Nyamenthi et al. 414 451 Inclusion: 18–46 years of age, Nurse case management (NCM) Urinalysis drug testing results, Proportion of participants in both
(2017); Zhang self-reported being homeless, + CM or standard education self-report of number of sexual groups tested positive for meth use
et al. (2018) GBM, stimulant use within the (SE) + CM for 16 weeks Follow- partners HAV/HBV Twinrix decreased from more than 70 % at
past 3 months, no self-reported up: 4 and 8-month Design: RCT vaccine completion baseline to less than 30 % at 8-months.
participation in substance abuse Proportion of participants in both
treatment in the past 30 days groups reported having sex with
multiple partners decreased from 38%
at baseline to 17.28% for the
NCM + CM participants and 21.59%
for the SE + CM participants at 8-
months. 85% of participants eligible to
receive vaccine in both groups
completed the three-dose vaccination
series during the intervention periods.
Reback et al. 131 Inclusion: at least 18 years of CM or TAU for 24-weeks Attendance and health- CM participants achieved greater
(2010, 2012); age, met DSM-IV criteria for Follow-up: 7, 9, and 12-month promoting total scores reductions in stimulant and alcohol use
Fletcher and substance dependence, non- Design: RCT calculated at the end of the (χ2 = 27.36, p < .01), especially
Reback treatment-seeking, homeless, intervention, individual urine methamphetamine (χ2 = 21.78,
(2013); self-reported sex with a man in drug analysis and breathalyzer p < .01) than control participants. CM
Fletcher et al. the previous 12 months results and aggregate Level 1 participants achieved greater increases
(2014) Exclusion: unable to scores in health-promoting behaviors
understand consent forms, (χ2 = 37.83, p < .01) than control
determined to have a more participants. Treatment responders
serious psychiatric condition were more likely to be Caucasian/
white than non-responders (76.5% vs.
46.8%; p < .05). Treatment responders
reported significantly fewer years of
lifetime cocaine use (2.5 [SD = 3.6] vs.
5.9 [SD = 7.7]; p < .05), fewer years of
lifetime meth use (3.8 [SD = 4.2] vs.
6.7 [SD = 6.7]; p < .05), and fewer
years of lifetime polysubstance use (7.8
[SD = 4.4] vs. 12.9 [SD = 8.5];
p < .05) than non-responders.
Treatment responders reported more
male sexual partners in the previous six
months (11.8 [SD = 16.7] vs. 5.5
[SD = 8.5]; p = .0587) and a greater
number of reported unprotected anal
intercourse events (24.9 [SD = 61.4]
vs. 2.5 [SD = 9.9]; p < .05) than
responders. ASPD status was
significantly positively associated with
meth abstinence (coefficient = 0.1
[SE = 0.05]; p < .05). Participants
with ASPD earned less per health-
promoting behavior than those without
ASPD ($10.21 [SD = &7.02] versus
$18.38 [SD = $13.60]; p < .01). Meth
use during the course of the
intervention was associated with
significant decreases in the number of
(continued on next page)

4
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

Table 1 (continued )
Author (year) N Inclusion/Exclusion Criteria Intervention Description Outcomes Relevant Findings

days between voucher redemptions

(IRR = 0.66; 95% CI = [.44− 0.99]).
Reback and Study 1: Inclusion: 18− 65 years of age, Study 1: CM, standard CBT, Average retention in weeks of Modified GCBT + CM produced fewer
Shoptaw 40 self-identified gay/bisexual CM + CBT, or GCBT Design: intervention, longest consecutive weeks of meth abstinence
(2014) Study 2: men, seeking treatment for RCT Study 2: GCBT or gay- consecutive abstinence period (− 0.44, CI: − 0.79, − 0.09) and fewer
46 methamphetamine use specific social support therapy verified by urine drug screen male sexual partners (− 0.36, CI: − 0.71,
Study 3: Design: RCT Study 3: results (in weeks), treatment − 0.02) than the first trial of GCBT.
171 participated in a modified effectiveness score, percent of Modified GCBT + CM produced more
GCBT + low-cost CM Design: urine samples negative for meth days of meth use (0.35, CI: 0.02, 0.68)
Pre-post (no control) metabolite, and number of days than the second trial of GCBT. Modified
of reported meth use in the GCBT + CM produced greater effects in
previous 30 days reducing the number of male sexual
partners (− 0.54, CI: − 0.89, − 0.19;
− 0.51, CI: − 0.84, − 0.18) at 26-week
follow-up.
Roll et al. (2013) 118 Inclusion: 18− 65 years of age, Standard psychosocial Drug use measured by: longest Compared to the standard treatment
met DSM-V criteria for meth treatment, standard duration of continuous condition, 1 Month CM participants
dependence, were willing and psychosocial abstinence during the treatment were almost 4 times more likely to
able to comply with study treatment + either 1 month, 2 phase, total percent of meth submit a negative meth UA (odds ratio
procedures and provide written months, or 4 months of CM urine analyses (UAs) indicating (OR) = 3.58, p < 0.05), 2 Month CM
informed consent. Exclusion: Follow-up: 6, 8, 10, and 12- abstinence during treatment participants were about 2.5 times more
had medical or psychiatric months Design: RCT phase (e.g., 100% and 80%), and likely to submit a negative meth UA
conditions that precluded safe proportion of negative meth UAs (OR = 2.55, p < 0.05) and 4 Month CM
study involvement, were unable submitted across the treatment participants were about 7.25 times
to comply with the research period (i.e., 3 samples per week more likely to submit a negative meth
requirements, any other for 16 weeks). UA (OR = 7.25, p < 0.05).
circumstances that would
interfere with study
participation.
Roll et al. 88 Inclusion: adults seeking 5 conditions with varying CM Total abstinence from meth as No significant differences were
(2006a) outpatient behavioral treatment reinforcement schedules for 12 assessed by mean number of observed between groups in either
for meth use disorders weeks Follow-up: none Design: metabolite-free urine samples mean total number of abstinences
RCT provided, continuous abstinence during treatment (F(4, 78) = 0.4696,
measured by longest period of p > 0.05) nor in the mean longest
uninterrupted abstinence, the period of continuous abstinence (F{4,
ability to initiate abstinence 78) = 0.7551, p > 0.05). Schedule 5
from meth as defined as the initiated abstinence quicker than
mean number of tests that Schedule 1 (q = 4.492) and Schedule 3
occurred in each condition prior (q = 5.241). Participants assigned to
to producing the first meth- Schedule 5 were less likely to relapse
negative test result, the ability to after 4 weeks of abstinence (Chi-
protect against relapse following square = 8.084 (3), p < .05).
a period of abstinence as
assessed by counting the number
of participants who relapsed
following a 4-week period of
abstinence
Roll et al. 113 Inclusion: had diagnoses of Treatment as usual (TAU) Retention as defined by the CM participants submitted significantly
(2006b) * either meth abuse or either Matrix Model, CBT, or number of weeks that elapsed more stimulant- and alcohol-negative
dependence relapse prevention depending between the first and last study samples (mean = 13.9, SD = 8.8) than
on clinic or TAU plus CM for 12 urine samples submitted and TAU participants (mean = 9.9,
weeks Follow-up: 3 and 6- whether the participant SD = 8.0) (t = 2.55, df = 110,
months Design: RCT completed the study Treatment p = 0.01). CM participants had a longer
participation was evaluated by mean period of documented
examining the number of continuous abstinence (approximately
counseling sessions attended 4.6 weeks, based on a mean of 9.3
during the 12-week period. Drug consecutive samples, SD = 9.2) than
use as measured by total number TAU participants (approximately 2.8
of stimulant-and alcohol- weeks, based on a mean of 5.6
negative samples submitted by consecutive samples, SD = 7.2)
each participant, test results for (t = 2.38, df = 110, p = 0.02). CM
stimulants and alcohol at each participants were more likely to submit
visit and follow-up, and longest negative urine samples than TAU
documented duration of participants (odds ratio = 2.04, 95%
sustained abstinence from CI = 1.19–3.49; χ2 = 6.80, df = 1,
stimulant drugs and alcohol for p = 0.009).
each participant.
Roll and 18 Inclusion: met DSM-IV criteria VBRT with reset or VBRT abstinence verified by urine Escalating-with-reset participants
Shoptaw for meth dependence without reset for 12 weeks drug analysis results submitted more total meth-free urine
(2006) Follow-up: none Design: RCT samples (80 % of all samples)
compared with escalating without-
reset participants (38% of all samples)
(t = 3.1, df = 16, p < 0.05).
Escalating-with-reset participants
achieved longer periods of continuous
abstinence (mean = 6.7 weeks,
(continued on next page)

5
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

Table 1 (continued )
Author (year) N Inclusion/Exclusion Criteria Intervention Description Outcomes Relevant Findings

SD = 3.2 weeks) than escalating-

without-reset participants (mean = 2.8
weeks, S. = 3.6 weeks) (t = 2.4,
df = 16, p < 0.05).
Shoptaw et al. 162 Inclusion: Gay and bisexual CM, cognitive behavioral Recent stimulant use measured Significant differences were observed
(2005)*; Ling men (GBM), aged 18–65, who therapy (CBT) + CM, CBT, by urine samples analyzed for in retention (F(3,158) = 3.78, p < .02),
Murtaugh were seeking outpatient culturally-tailored cognitive metabolites of meth and longest consecutive period of urine
et al. (2013) behavioral drug abuse behavioral therapy (GCBT) for cocaine. Results from urine drug samples negative for meth metabolites
treatment for meth dependence 16-weeks Follow-up: 6 and 12- screening were compiled using (F(3,158) = 11.80, p < .001), and TES
at a research clinic Exclusion: months Design: RCT three indices: Treatment by condition during treatment (F
had medical or psychiatric Effectiveness Score (TES); (3,158) = 7.35, p < .001) with CM and
conditions that precluded safe percent of urine samples that CBT + CM conditions performing
study involvement, were unable were metabolite-free; and better than standard CBT. GCBT
to comply with the research longest period of consecutive participants showed a reduction in
requirements, or if their meth metabolite-free urine samples unprotected receptive anal intercourse
dependence required more (in days). Frequency and (URAI) during the first 4 weeks of
intensive intervention than magnitude of CM purchases, treatment (X2 = 6.75, p < .01).
outpatient treatment. purchase type as either hedonic/ Overall reductions in outcomes were
utilitarian and consumable/ sustained to 1-year, but group
durable based on physical differences were not sustained at
records of purchases follow-up. Participants who redeemed
CM earnings at any visit (“spenders”)
were significantly more likely to
produce stimulant-negative urine
samples in the subsequent visit than
those who did not redeem (“savers”)
1.011*[1.005, 1.017], Z = 3.43,
p < .001.
Shoptaw et al. 229 Inclusion: 18− 65 years of age, Sertraline + CM, sertraline- Meth use (urine drug screening Participants in the sertraline-only
(2006)* met DSM-IV criteria for MA only, placebo + CM, and and self-reported days of use), condition provided more metabolite-
dependence or abuse placebo-only conditions for 12 retention (length of stay), drug positive urine samples compared to the
Exclusion: had a primary weeks Follow-up: none Design: craving (visual analogue scale), other conditions over the trial (GEE
medical condition that might RCT and mood symptoms (Beck model: χ2(1) = 5.02, p < 0.05). Fewer
interfere with safe study Depression Inventory) sertraline-only participants (25.4%)
participation, current achieved the outcome criterion of at
pharmacotherapy for which least 3 weeks of consecutive meth
concurrent treatment with a metabolite-free urine samples
selective serotonin reuptake compared to participants in other
inhibitor would be conditions (sertraline + CM = 42.6%,
contraindicated, a psychiatric placebo + CM = 51.9%, and placebo-
condition identified by the SCID only = 41.8%; χ2 (3) = 8.6,
that required pharmacological p = 0.035). More participants
or behavioral intervention, or receiving CM (47.0%) achieved three
SCID-diagnosed dependence on consecutive weeks of meth abstinence
opioids, cocaine, alcohol, or than those not receiving CM (33.3%; χ2
benzodiazepines. (1) = 4.4, p = 0.036) when the four
study conditions were collapsed to
participants receiving CM versus those
not receiving CM.
*
Indicates studies included in previous review by Lee and Rawson (2008).

management studies were considered including those selected for pre­ disappeared. Reback and Shoptaw (2014) compared the outcomes of
vious reviews. All study designs (included pilots), comparator groups, Shoptaw et al. (2005) and two other studies using a meta-analysis. Of the
and outcomes were considered. Outcomes reported here include drug three studies included in the meta-analysis, two included a contingency
abstinence, retention in treatment, attendance/treatment engagement, management component. The second study did not include a contin­
sexual risk behavior, drug craving, mood/affect, intervention accept­ gency management component, but rather evaluated a
ability, and treatment response predictors. culturally-tailored cognitive behavior therapy and a gay-specific social
support therapy (Shoptaw et al., 2008). Participants in the third study all
3. Results received a modified culturally-specific cognitive behavioral therapy
(reduced from 16 weeks to 8 weeks) in conjunction with a low-cost
3.1. Randomized controlled trials contingency management intervention. The modified treatment and
the original culturally-specific cognitive behavioral therapy both pro­
3.1.1. Drug abstinence duced reductions in methamphetamine use (Shoptaw et al., 2005), but
Shoptaw et al. (2005) found that methamphetamine dependent men the modified treatment produced fewer consecutive weeks of metham­
who have sex with men who received contingency management or phetamine abstinence than the original.
contingency management + cognitive behavioral therapy achieved Roll et al. (2006b) reported that participants diagnosed with meth­
greater Treatment Effectiveness scores (proportion of samples negative amphetamine dependence or abuse who received treatment as usual­
for cocaine and methamphetamine metabolites) and longer durations of + contingency management submitted significantly more stimulant-
abstinence than participants who received only cognitive behavioral and alcohol-negative samples and achieved longer periods of continuous
therapy. These reductions in methamphetamine use were sustained abstinence than participants who received only treatment as usual. This
through the 12-month follow-up; however, group differences difference was no longer apparent at the 3- and 6-month post-treatment

6
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

Table 2
Summary of Pre-post Studies.
Author (year) N Inclusion/Exclusion Criteria Intervention Description Outcomes Relevant Findings

Gómez et al. (2018) 131 Inclusion: meth using men who Participated in community- abstinence verified by urine drug Participants submitted a median of
have sex with men (MSM) based CM program (PROP) + analysis results 22 out of 36 samples
individual and drop-in group (IQR = 10− 34) nonreactive to
counseling for 12 weeks meth, which equates to over 7
Follow-up: none Design: Pre- weeks abstinent of the 12 week
post (no control) treatment duration
Landovitz et al. 53 Inclusion: self-identified as MSM, Participated in a CM Urinalysis drug testing results, HIV Mean number of days (of the past
(2012) were at least 18 years of age, were intervention for 12 weeks and STI testing results, medication 30) of meth use decreased from 4.8
HIV uninfected on rapid HIV ELISA Follow-up: 3-month Design: adherence via self-report and pill to 1.6 (p < 0.001) and mean
testing, self-reported meth use Pre-post (no control) counts number of uses per day decreased
within the previous 30 days, and from 5.3 to 1.1 (p < 0.001) from
reported unprotected anal baseline to 3-month follow-up. The
intercourse (UAI) with an HIV- percentage of participants who
positive or HIV-serostatus- provided metabolite-free urine
unknown partner in the previous samples increased from 71.7 at
90 days. baseline to 89.47% at 3-month
follow-up (p < 0.05). Meth
abstinence during CM treatment
increased PEP adherence (2% [95%
CI + 1–3%] per metabolite-free
sample provided), and increased
the odds of PEP course completion
(OR 1.17, 95% CI 1.04–1.31).
Okafor et al. 30 Inclusion: met DSM-5 criteria for Participated in a pilot CM Therapeutic response defined as 68 % of participants submitted ≥23
(2019); meth use disorder, provided a trial for 8 weeks Follow-up: abstinence from methamphetamine meth-negative urine samples and
Krishnamurti urine sample that tested positive none Design: Pre-post (no (≥23 of 24 possible were classified as responders. Fewer
et al. (2020); for meth during screening, were control) methamphetamine-negative urine responders reported monthly
Lake et al. (2020) aged 18–45 years, had a good samples). Urinalysis drug testing household income ≥25 000+ South
understanding of English results, records of purchases African Rand (ZAR; ~USD $1880;
Exclusion: currently in treatment vs. ZAR < 25 000) than non-
for addiction to a substance other responders (15.8% vs. 63.6%;
than stimulants; met DSM-5 P = 0.007). Responders had higher
criteria for a substance-related median years of education (12 vs.
diagnosis other than meth, 10; Kruskal–Wallis χ2 = 4.25,
tobacco, marijuana; current use of DF = 1, P = 0.039) and lower
a prescribed psychoactive median body mass index than non-
medication; inability to attend ≥ 4 responders (19 vs. 24;
visits during a 2-week screening Kruskal–Wallis χ2 = 6.84,
period or to complete screening P = 0.008). Participants who spent
measures; physical or mental earnings at a previous visit
illness that would require (“spenders”) were more likely to
intervention, alter brain imaging produce stimulant-negative urine
findings, or interfere with safe samples at subsequent visit than
study participation; pregnancy; those who did not (“savers”)
claustrophobia; presence of metal [OR = 1.23, CI = 1.08− 1.53,
prostheses, cardiac pacemakers or p = .002]. Partial responders
metal clips that are incompatible showed a greater preference for
with the magnetic resonance large, immediate rewards over
imaging environment; HIV smaller, short-term but larger long-
seropositive status, previous head term rewards and long-term losses
injury. than healthy controls [p = 0.038, g
= -0.77 (-1.09: -0.44)].
Shoptaw et al. 111 Inclusion: men who have sex with Participated in community- acceptability (number of Participants produced an average of
(2006) men with recent meth use based CM program (PROP) enrollments/ time), impact (clinical 15 (SD = 13.5) meth metabolite-
for 12 weeks Follow-up: none response to treatment and cost- free urine samples of the possible 36
Design: Pre-post (no control) effectiveness as cost per patient samples over the 12-week
treated) treatment period. 52% of
participants achieved 12
metabolite-free urine samples;
17%, 24 metabolite-free samples;
and 8.1%, all 36 metabolite-free
samples. 60% of participants
completed 4 weeks of treatment;
48%, 8 weeks and 30%, 12 weeks.
Strona et al. (2006) 178 Inclusion: met criteria for meth Participated in community- abstinence verified by urine drug Of submitted samples, 96 % were
dependency, reported using meth based CM program (PROP) analysis results, self-report of number negative for meth. A significant
in last 7 days, were not enrolled or for 12 weeks Follow-up: none of sex partners reduction in number of self-
active in a drug treatment Design: Pre-post (no control) reported sex partners was achieved
program, participated in a 15-min (p < .05).
intake session, agreed to submit
thrice-weekly urine samples for 12
weeks

7
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

follow-ups. were reported; however, participants assigned to the contingency

Roll et al. (2006a) reported that five different schedules of rein­ management + positive affect condition self-reported less stimulant use.
forcement for contingency management incentives produced compara­ Chudzynski et al. (2015) reported that participants seeking treat­
ble mean total number of negative samples and mean longest duration of ment for methamphetamine use disorders who received contingency
abstinence for participants seeking treatment for methamphetamine use management submitted more negative samples than participants who
disorders. However, a schedule developed by Higgins et al. (1994) received standard care. Amongst contingency management participants,
initiated abstinence more rapidly than the two other schedules and there were no differences in the number of negative samples associated
offered better protection against relapse than all other schedules. with different frequencies of reinforcement.
Roll and Shoptaw (2006) found that methamphetamine dependent Nyamathi et al. (2017) and Zhang et al. (2018) assigned homeless
participants assigned to an escalating-with-reset contingency manage­ stimulant-using men who have sex with men and transgender women to
ment schedule condition submitted more methamphetamine negative either a Nurse Case Management + contingency management or to a
samples and achieved longer periods of abstinence than participants in Standard Education + contingency management condition. However,
the escalating-without-reset condition. Nyamathi et al. (2017) excluded transgender women from their analysis
Shoptaw et al. (2006a) reported that significantly more participants given the distinct needs of this population. In both studies, both condi­
who met criteria for methamphetamine abuse or dependence assigned to tions produced a reduction in the number of participants who tested
either a contingency management + sertraline or contingency man­ positive for methamphetamine.
agement + placebo condition achieved three weeks of abstinence than
participants assigned to either sertraline- or placebo-only conditions. 3.1.2. Retention
Interestingly, participants in the contingency management groups ach­ Shoptaw et al. (2005) reported that contingency management par­
ieved these results despite earning a considerably smaller proportion of ticipants retained for longer periods than cognitive behavioral therapy
the incentives than had been observed in earlier contingency manage­ participants. Reback et al. (2010) found that contingency management
ment studies. participants submitted a higher proportion of samples, and Roll et al.
Reback et al. (2010) found that homeless, substance-using men who (2013) found that retention rates were significantly higher in the
have sex with men in a contingency management condition who 4-month contingency management group than the standard treatment
received incentives for abstinence, attendance, and completion of group. However, Roll et al. (2006b) and Menza et al. (2010) found
health-promoting behaviors submitted more samples negative for similar retention rates between conditions.
methamphetamine, amphetamines, PCP, and cocaine and a blood
alcohol <.05 than participants in a control group who received in­ 3.1.3. Attendance/Treatment engagement
centives only for attendance. Contingency management participants Six studies reported effects of contingency management on atten­
retained reductions in drug use compared to control participants dance and/or treatment engagement. Shoptaw et al. (2005) found that
through 9- and 12- month follow-ups. contingency management + cognitive behavioral therapy participants
Menza et al. (2010) did not find group differences in the submission attended significantly more therapy sessions over standard cognitive
of negative samples during intervention for methamphetamine-using behavioral therapy participants and achieved significantly higher pay­
men who have sex with men randomly assigned to receive contin­ outs than contingency management-only participants. Chudzynski et al.
gency management or referral to community resources. During inter­ (2015) found that participants in the intermittent predictable contin­
vention and at follow-up the contingency management group was gency management condition were more likely to attend treatment than
significantly more likely to self-report weekly or daily methamphet­ those in the standard treatment condition. Corsi et al. (2019) found that,
amine use and using >8 quarters of methamphetamine. among the subset of participants who had a regular sex partner, those
In Corsi et al. (2012) and Corsi et al. (2019), non-transient men who assigned to the contingency management + strengths-based case man­
have sex with men were randomly assigned to a contingency manage­ agement condition were more likely to attend at least one session than
ment group or a contingency management + strengths-based case those in in the contingency management-only condition. Zhang et al.
management group. Both studies showed improvements in participants (2018) found that the majority of participants eligible to receive the
over time, but neither study indicated a significant difference between HAV/HBV vaccination completed the vaccine series, independent of
groups on the main outcome variables. In the Corsi et al. (2012) study, a whether they were assigned to the Nurse Case Manage­
significant reduction in the proportion of amphetamine- and ment + contingency management or the Standard Educa­
methamphetamine-positive samples from baseline to follow-ups were tion + contingency management conditions. Roll et al. (2006b) found
reported in both conditions. no effects of treatment on counseling attendance. Reback et al. (2010)
Roll et al. (2013) reported that for methamphetamine dependent also reported no differences between conditions in attendance at study
participants, consecutive days of abstinence and the number of partici­ visits or service program activities; however, contingency management
pants who remained 100 % or 80 % abstinent increased as contingency participants completed significantly more health-promoting behaviors.
management duration increased (no contingency management, 1-, 2-, or
4-months). However, treatment attendance and consecutive days of 3.1.4. Sexual risk behavior
methamphetamine abstinence were positively correlated with duration Three studies reported reductions in risky sex. Shoptaw et al. (2005)
of contingency management making it difficult to separate the effects of and Menza et al. (2010) found that participants across conditions re­
contingency management on abstinence and attendance. Specifically, ported a decrease in unprotected anal intercourse at the end of treatment
there is an interpretive difficulty that arises from imputing missing from baseline; these reductions were maintained through follow-up.
samples as positive, especially when a direct effect of the intervention is Corsi et al. (2012) found that at the 4-month follow-up strengths-­
to reinforce attendance (Silverman et al., 2013). based case management + contingency management participants re­
Carrico et al. (2015) randomly assigned men who have sex with men ported greater reductions in sex risk behaviors including any sex in the
to receive contingency management only or contingency management last 30 days, unprotected sex, sex under the influence, and sex for drugs
plus a positive affect intervention. They reported no differences in the or money than contingency management-only participants. However, at
number of methamphetamine negative samples submitted or the mean the 8-month follow-up the effect of treatment was reversed for sex under
duration of continuous abstinence between groups. In both groups, the the influence and sex for drugs or money.
majority of submitted samples were negative. Carrico et al. (2018) found Four studies reported reductions in the number of sexual partners.
similar results with a sample of HIV-positive participants. No differences Shoptaw et al. (2005) reported reductions in the number of sexual
between groups in the number of methamphetamine negative samples partners across conditions. Nyamathi et al. (2017) and Zhang et al.

8
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

(2018) found that fewer participants in both groups (i.e., Nurse Case follow-up.
Management + contingency management and Standard Educa­ Fletcher et al. (2014) and Ling Murtaugh et al. (2013) evaluated
tion + contingency management) reported sex with multiple partners whether the frequency of voucher redemption was associated with
from baseline to the 4-month follow-up. Reback et al. (2014) found that substance use. Fletcher et al. (2014) found that participants who used
a modified culturally specific cognitive behavioral ther­ methamphetamine throughout the intervention had less time between
apy + contingency management intervention produced greater re­ their voucher exchanges. Contrarily, Ling Murtaugh et al. (2013) found
ductions in number of male sexual partners at the end of treatment and that participants who exchanged vouchers during a visit were more
at follow-up than culturally specific cognitive behavioral therapy -only likely to produce a stimulant-negative sample at the next visit than
interventions. participants who accumulated their vouchers.
Additional associations with treatment response were identified in
3.1.5. Drug craving three studies. Reback et al. (2012) found that responders reported more
Shoptaw et al. (2006) did not find any interaction between treatment recent sexual partners and unprotected anal intercourse events. Corsi
and methamphetamine craving, but they did find an interaction be­ et al. (2019) found that belonging to a class of participants who were in a
tween time and craving. Carrico et al. (2018) found that participants in couple, had no history of sexual abuse, and reported less methamphet­
the positive affect group reported less intense cravings prior to the final amine at baseline was a predictor for negative samples. Within this class
three positive affect sessions and at the end of treatment. of participants, those in the strengths-based case manage­
ment + contingency management submitted more negative samples.
3.1.6. Mood/Affect Fletcher and Reback (2013) found that participants diagnosed with
Carrico et al. (2015) found that participants in the positive Antisocial personality disorder (ASPD) submitted more negative sam­
affect + contingency management condition reported increases in pos­ ples than participants without ASPD regardless of condition and that
itive affect at the end of treatment and participants in the contingency ASPD participants completed a similar amount of health-promoting
management-only group reported decreases in negative affect; however, behaviors but earned fewer vouchers for health-promoting behaviors
these differences were not maintained through follow-up. Carrico et al. indicating that they completed smaller magnitude behaviors.
(2018) found that positive-affect + contingency management partici­
pants reported an increase in positive affect prior to sessions 3 and 5. 3.2. Pre-post studies
Shoptaw et al. (2006) did not find an interaction between treatment and
depressive symptoms. 3.2.1. Drug abstinence
Strona et al. (2006); Shoptaw et al. (2006b), and Gómez et al. (2018)
3.1.7. Acceptability evaluated the Positive Reinforcement Opportunity Project (PROP), a
Corsi et al. (2012) found that participants rated the contingency contingency management intervention for reducing methamphetamine
management intervention positively and that participants in the use in men who have sex with men that was implemented by the San
strengths-based case management + contingency management condi­ Francisco Department of Public Health. Participants in the Shoptaw
tion rated the testing schedule more positively and barriers to atten­ et al. (2006b) study submitted an average of 15 of a possible 36
dance and participation less negatively than contingency methamphetamine-negative samples, and participants in the Gómez
management-only participants. Zhang et al. (2018) found that vacci­ et al. (2018) study submitted a median of 22 of a possible 36 negative
nation as a result of both Nurse Case Management + contingency samples over the course of the intervention, equating to over 7 weeks of
management and Standard Education + contingency management was abstinence. Strona et al. (2006) reported that 96 % of all submitted
more cost-effective than HBV treatment and that Standard Educa­ samples were methamphetamine-negative. This suggests that the large
tion + contingency management was more cost-effective and required majority of PROP participants who continued to use methamphetamine
half of the staff time than did Nurse Case Management + contingency during the Strona et al. (2006) study did not submit samples.
management. Landovitz et al. (2012) evaluated a contingency management inter­
vention among methamphetamine using men who have sex with men
3.1.8. Treatment response predictors who had recently reported a high-risk sexual or injection drug exposure
HIV serostatus was associated with treatment response in three and were, therefore, initiated on Post-Exposure Prophylaxis (PEP). Over
studies. Menza et al. (2010) found that HIV-positive participants were the course of the intervention, participants reported a reduction in
more likely to attend the 24-week study visit. Reback et al. (2010) found methamphetamine use and submitted more methamphetamine negative
that within the CM condition, HIV-positive participants accomplished samples.
more health-promoting behaviors than HIV-negative participants and Okafor et al. (2019) conducted a secondary analysis of an 8-week
across conditions, HIV-negative participants submitted more amphet­ pilot contingency management intervention trial in South Africa and
amine, methamphetamine, and Level 1 composite negative samples than reported that 68 % of participants submitted a majority of metham­
HIV-positive participants. Nyamathi et al. (2017) reported that being phetamine negative samples and 61 % remained abstinent throughout
HIV-positive at baseline was a predictor for stimulant use at follow-up. the intervention.
Ethno-racial identity was associated with treatment response in three
studies. Reback et al. (2010) reported that Caucasian participants 3.2.2. Retention
accomplished more health-promoting behaviors than African American Strona et al. (2006) and Shoptaw et al. (2006b) reported 35 % and 30
and Latino participants. and Caucasian participants submitted more % retention rates at the end of treatment for PROP participants,
Level 1 composite negative samples than all other ethnic groups. Reback respectively. Okafor et al. (2019) reported a 93 % treatment completion
et al. (2012) found that responders were more likely to be Caucasian. rate.
Nyamathi et al. (2017) found that “other” race/ethnicity was inversely
related to stimulant-use at follow-up. 3.2.3. Attendance/treatment engagement
Reback et al. (2012) and Nyamathi et al. (2017) found associations Landovitz et al. (2012) reported that the likelihood of PEP adherence
between a history of drug use and treatment response. In the Reback and course completion increased with the number of
et al. (2012) study, responders reported fewer years of lifetime use of methamphetamine-negative samples submitted.
cocaine, methamphetamine, as well as polysubstance use. Nyamathi
et al. (2017) reported that recent injection drug use and testing positive 3.2.4. Acceptability
for amphetamines at baseline were predictors for stimulant use at Strona et al. (2006); Shoptaw et al. (2006b), and Gómez et al. (2018)

9
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

reported treatment effectiveness comparable to that of treatment management in combination with another treatment, including nurse
research programs. Additionally, Strona et al. (2006) and Shoptaw et al. case management (Nyamathi et al., 2017; Zhang et al., 2018), cognitive
(2006b) reported retention rates comparable to research studies and behavioral therapy (Shoptaw et al., 2005), culturally-tailored cognitive
lower cost of treatment than other available treatments. behavioral therapy (Reback and Shoptaw, 2014), pharmacotherapy
Krishnamurti et al. (2020) interviewed participants following an (Shoptaw et al., 2006a), strengths-based case management (Corsi et al.,
8-week contingency management intervention in South Africa and 2012, 2019), and a positive affect intervention (Carrico et al., 2015,
found that a majority of participants (12/17) viewed incentives for 2018). In the majority of these studies, treatment outcomes related to
abstinence positively and most participants (15/17) viewed them as methamphetamine use were not improved by the addition of another
rewards which they were proud of earning. treatment and one study found that it was more cost-effective to deliver
standard contingency management (Zhang et al., 2018). However,
3.2.5. Sexual risk behaviors Carrico et al. (2018) found that participants who also received a positive
Strona et al. (2006) and Landovitz et al. (2012) observed a reduction affect intervention self-reported less stimulant use and methamphet­
in the self-reported number of sex partners over the course of treatment. amine craving than those who did not; and Reback and Shoptaw (2014)
Landovitz et al. (2012) also observed a reduction in the number of un­ found that while the addition of a culturally-tailored cognitive behav­
protected anal intercourse events. ioral therapy did not improve outcomes related to methamphetamine
use, it produced reductions in reported male sex partners.
3.2.6. Treatment response predictors Contingency management has been shown to be efficacious in
Okafor et al. (2019) reported that responders to treatment had lower reducing methamphetamine use for special populations that are
monthly income, more years of education, and lower body mass index impacted by methamphetamine use, including men who have sex with
than non-responders. Lower body mass index was also correlated with men (Carrico et al., 2015, 2018; Gómez et al., 2018; Landovitz et al.,
lower monthly income, indicating that both may be a product of a 2012; Nyamathi et al., 2017; Reback et al., 2010; Reback and Shoptaw,
general relationship between socioeconomic status and response to 2014; Shoptaw et al., 2006a; Strona et al., 2006; Zhang et al., 2018).
treatment. This may be especially important as methamphetamine is associated
Krishnamurti et al. (2020) found that participants who exchanged with risky sexual behavior that significantly contributes to the trans­
vouchers at one study session were more likely to submit negative mission of HIV and other sexually transmitted infections among men
samples at subsequent study sessions than participants who accumu­ who have sex with men (Colfax et al., 2004; Marshall et al., 2011).
lated vouchers, replicating the findings of Ling Murtaugh et al. (2013). Studies found a variety of collateral effects of contingency management
Lake et al. (2020) assessed the decision-making of partial responders interventions for methamphetamine use disorder, including fewer sex
(participants who submitted one or more methamphetamine positive or partners (Nyamathi et al., 2017; Reback and Shoptaw, 2014; Strona
missing sample) compared to healthy controls using the Iowa Gambling et al., 2006; Zhang et al., 2018), fewer instances of unprotected anal
Task. While all participants show a preference for larger more imme­ intercourse (Landovitz et al., 2012), increased completion of an
diate rewards, partial responders made more choices for large, imme­ HAV/HBV vaccine series (Zhang et al., 2018) and adherence to PEP
diate rewards over smaller short-term rewards relative to healthy (Landovitz et al., 2013).
controls. Several studies identified possible predicters of treatment outcomes.
Contingency management interventions were least effective for partic­
4. Discussion ipants who reported a longer history of drug use (Reback et al., 2012) or
reported more methamphetamine use during baseline (Corsi et al.,
The current literature strongly supports the use of contingency 2019). Other possible predictors were race with Caucasian participants
management as an intervention for methamphetamine use and provides being most likely to respond to treatment (Reback et al., 2010, 2012)
further evidence for the use of contingency management as a treatment and HIV status with HIV negative participants responding more posi­
for substance use more broadly. The present review provides the most tively to treatment (Nyamathi et al., 2017; Reback et al., 2010). Okafor
complete account to date regarding the broad beneficial effects of con­ et al. (2019) found that responders to treatment were more likely to
tingency management in the treatment of methamphetamine use dis­ have a lower monthly income, more years of education, and a lower
order. This includes its effects on drug abstinence, drug craving, body mass index. Lake et al. (2020) found that full responders were
treatment engagement and retention, earning and spending of in­ similar to healthy controls, but partial responders preferred larger more
centives, mood/affect, and sexual risk behaviors, as well as intervention immediate rewards to a greater extent than healthy controls. Taken
acceptability from the perspective of patients. The present review also together, these results suggest that greater problem severity and
provides detailed summary of intervention parameters and predictor increasingly complex constellations of medical, social, and psychologi­
variables correlated with the outcome of contingency management cal challenges may reduce the efficacy of contingency management
intervention in methamphetamine use disorder patients. intervention. However, it is not therefore appropriate to conclude that
contingency management should be withheld from a person with greater
4.1. Broad effects and correlates of success problem severity. Rather, it may be necessary to provide even more
intensive, higher magnitude contingency management intervention
Of the 27 studies included in this review, only one (Menza et al., accompanied by a broader array of psychosocial services to especially
2010) found that contingency management did not effectively reduce refractory patients.
methamphetamine use. This may have been the result of participants Three studies included analyses to determine whether then spending
completing twice-weekly urine screenings as opposed to thrice-weekly, behavior of contingency management participants had implications for
because longer inter-test intervals can allow methamphetamine use to treatment outcomes. Two studies (Ling Murtaugh et al., 2013; Krish­
go undetected and potentially compromise the contingent reinforcement namurti et al., 2020) found that participants who spent earned vouchers
of abstinence. were more likely to submit methamphetamine negative samples at
One study reviewed here evaluated the effects of contingency man­ subsequent sessions. However, Fletcher et al. (2014) observed that
agement duration and found that longer intervention (up to 4 months) participants who saved vouchers were more likely to submit negative
produced better outcomes (Roll et al., 2013). While no other study here samples.
examined intervention duration, included study interventions ranged The beneficial effects of contingency management for methamphet­
from 2 to 4 months and were effective. amine are also apparent when it is implemented in community pro­
Several of the reviewed studies were evaluations of contingency grams. Three studies reviewed here (Gómez et al., 2018; Shoptaw et al.,

10
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

2006b; Strona et al., 2006) evaluated the effectiveness of a barriers (Kurti et al., 2015), but cost-benefit analyses, and intervention
community-based program in San Francisco, the Positive Reinforcement refinements that improve cost-benefit are clearly priorities for future
Opportunity Project (PROP). Outcomes similar to those observed in research if the behavior of decision makers is to be changed in favor of
research treatment programs were observed for PROP participants as the adoption of contingency management. Robust post-treatment ef­
well, and the intervention was found to be acceptable to participants and fects, in combination with clear evidence of cost-effectiveness and a
feasible to implement. method of intervention delivery that reduces effort and obviates
expertise would almost certainly shift the decision making of potential
4.2. Comparison to the treatment of other substance use disorders (SUDs) adopters. Although it may require substantial time and effort on the part
of substance abuse treatment researchers to produce an ideal contin­
The effects of contingency management on methamphetamine gency management intervention, optimism is warranted. The Depart­
abstinence and treatment retention and engagement are not surprising, ment of Veterans Affairs has implemented contingency management on
given that studies involving other SUDs have shown similar effects. For a national scale (Petry et al., 2014), and incentives to promote health
example, contingency management has long been known to reduce use behaviors are becoming increasingly accessible to Medicaid recipients
of alcohol, cannabis, cocaine, nicotine, and opioids (Lussier et al., 2006). (Vulimiri et al., 2019). Given the persistence of the problem itself, the
Similarly, the effectiveness of contingency management in promoting lack of pharmacotherapies, contingency management’s solid empirical
treatment attendance (Petry et al., 2018) and medication adherence and conceptual foundations, and the substantial progress that has been
(DeFulio and Silverman, 2012) in SUD populations is well-established. made in intervention development and dissemination over the last
From a conceptual perspective, the view that drug taking is behavior decade, the widespread provision of contingency management services
that is sensitive to its consequences is a cornerstone of behavioral to people with methamphetamine use disorders appears realistic.
pharmacology (Thompson and Johanson, 1981). The potential for con­
tingency management intervention to have broad beneficial effects
4.4. Limitations
beyond the behaviors specifically targeted by the intervention has also
been occasionally noted in studies targeting other SUDs (e.g., Silverman
There are several limitations of this review. First, because of the
et al., 1998). However, contingency management for methamphetamine
variety of intervention combinations and outcome measure variables, it
use disorder appears to be unique in terms of the level of support for the
may be hard to make comparisons between studies. Furthermore, this
collateral reduction is sexual risk behaviors. We speculate that this is
review, like any other, is subject to publication bias. This review only
because contingency management studies involving methamphetamine
included published studies and as such may not reveal a full picture of
use disorder have been more likely to include assessment of sexual risk
contingency management.
behavior as part of the research plan, and perhaps because these studies
have focused on populations with high baseline levels of sexual risk
5. Conclusion
behaviors.

Contingency management is broadly effective in reducing metham­

4.3. Gaps in the literature
phetamine use and promoting attendance to recovery-related appoint­
ments in people with methamphetamine use disorders, and appears to
There are two key gaps in the literature regarding contingency
produce broad beneficial effects beyond the behaviors targeted for
management in the treatment of methamphetamine use disorders. These
intervention. The available evidence supports the adoption of contin­
gaps are related.
gency management in outpatient treatment programs that serve meth­
The first gap is the need for interventions that support long-term
amphetamine users. Additional research designed to improve post-
drug abstinence. The state of affairs is well captured by a recent meta-
treatment outcomes and enhance intervention adoptability is war­
analysis, which showed that contingency management interventions
ranted, but immediate widespread implementation of VBRT in a manner
produced significant improvements in drug abstinence at three months
consistent with published studies is a higher priority.
post-treatment, but not at six months post-treatment (Sayegh et al.,
2017). However, the situation may be better than the meta-analysis
implies. In a review of contingency management studies published be­ Contributors
tween 2009 and 2014, the evidence for post-treatment effects was
summarized. During this period, 47 % of contingency management Hayley D. Brown, principal author, contributed to all sections of the
studies reported post-intervention effects, and 62 % of these studies (i.e., manuscript, conducted review operations as described in the manu­
29 % of the total) reported significant post-intervention increases in script, and created the table. Anthony DeFulio, second author, contrib­
drug abstinence (Davis et al., 2016). Long-term outcomes would likely uted to all sections of the manuscript, and conducted review operations
be much better if the drug abstinence contingencies were simply as described in the manuscript.
maintained, and at least one practical method for achieving this has
been devised (Silverman et al., 2012). Nevertheless, a method for sup­ Role of funding source
porting long-term abstinence that is both practical and scalable remains
an important challenge in the treatment of methamphetamine use dis­ Nothing declared.
order, and SUDs in general.
The second key gap is the adoption of contingency management by Declaration of Competing Interest
substance abuse treatment providers. Based on the behavioral principles
that underlie contingency management, the decision to adopt contin­ The authors report no declarations of interest.
gency management intervention is itself behavior that is a function of
the prior experience of the person making the decision and the contin­
References
gencies operating on the person at the moment of the decision. This
interpretation is consistent with studies of the barriers to adoption of AshaRani, P.V., Hambali, A., Seow, E., Jie, W., Tan, J.H., Subramaniam, M., 2020. Non-
contingency management, which have found that intervention costs, pharmacological interventions for methamphetamine use disorder: a systematic
effort, exposure to contingency management, and training are critical review. Drug Alcohol Depend. 212, 108060 https://doi.org/10.1016/j.
drugalcdep.2020.108060.
factors (Kirby et al., 2006). Post-treatment effects are also a key issue Baker, A., Lee, N.K., 2003. A review of psychosocial interventions for amphetamine use.
(Petry, 2010). Technology may play a role in overcoming many of these Drug Alcohol Rev. 22, 323–335. https://doi.org/10.1080/0959523031000154472.

11
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

Carrico, A.W., Gómez, W., Siever, M.D., Discepola, M.V., Dilworth, S.E., Moskowitz, J.T., Lussier, J.P., Heil, S.H., Mongeon, J.A., Badger, G.J., Higgins, S.T., 2006. A meta-analysis
2015. Pilot randomized controlled trial of an integrative intervention with of voucher-based reinforcement therapy for substance use disorders. Addiction 101,
methamphetamine-using men who have sex with men. Arch. Sex. Behav. 44, 192–203. https://doi.org/10.1111/j.1360-0443.2006.01311.x.
1861–1867 https://dx.doi.org/10.1007%2Fs10508-015-0505-5. Marshall, B.D., Wood, E., Shoveller, J.A., Patterson, T.L., Montaner, J.S., Kerr, T., 2011.
Carrico, A.W., Gόmez, W., Jain, J., Shoptaw, S., Discepola, M.V., Olem, D., Lagana- Pathways to HIV risk and vulnerability among lesbian, gay, bisexual, and
Jackson, J., Andrews, R., Neilands, T.B., Dilworth, S.E., Evans, J.L., Woods, W.J., transgendered methamphetamine users: a multi-cohort gender-based analysis. BMC
Moskowitz, J.T., 2018. Randomized controlled trial of a positive affect intervention Public Health 11, 20. https://doi.org/10.1186/1471-2458-11-20.
for methamphetamine users. Drug Alcohol Depend. 192, 8–15. https://doi.org/ Melin, G.L., Götestam, K.G., 1973. A contingency management program on a drug-free
10.1016/j.drugalcdep.2018.07.029. unit for intravenous amphetamine addicts. J. Behav. Ther. Exp. Psychiatry 4,
Chudzynski, J., Roll, J.M., McPherson, S., Cameron, J.M., Howell, D.N., 2015. 331–337. https://doi.org/10.1016/0005-7916(73)90003-7.
Reinforcement schedule effects on long-term behavior change. Psychol. Rec. 65, Menza, T.W., Jameson, D.R., Hughes, J.P., Colfax, G.N., Shoptaw, S., Golden, M.R., 2010.
347–353. https://doi.org/10.1007/s40732-014-0110-3. Contingency management to reduce methamphetamine use and sexual risk among
Colfax, G., Vittinghoff, E., Husnik, M.J., McKirnan, D., Buchbinder, S., Koblin, B., men who have sex with men: a randomized controlled trial. BMC Public Health 10,
Celum, C., Chesney, M., Huang, Y., Mayer, K., Bozeman, S., Judson, F.N., Bryant, K. 774. https://doi.org/10.1186/1471-2458-10-774.
J., Coates, T.J., 2004. Substance use and sexual risk: a participant-and episode-level Nyamathi, A., Reback, C.J., Shoptaw, S., Salem, B.E., Zhang, S., Yadav, K., 2017. Impact
analysis among a cohort of men who have sex with men. Am. J. Epidemiol. 159, of tailored interventions to reduce drug use and sexual risk behaviors among
1002–1012. https://doi.org/10.1093/aje/kwh135. homeless gay and bisexual men. Am. J. Mens Health 11, 208–220. https://doi.org/
Corsi, K.F., Lehman, W.E., Min, S.J., Lance, S.P., Speer, N., Booth, R.E., Shoptaw, S., 10.1177/1557988315590837.
2012. The feasibility of interventions to reduce HIV risk and drug use among Okafor, C.N., Stein, D.J., Dannatt, L., Ipser, J., van Nunen, L.J., Lake, M.T.,
heterosexual methamphetamine users. J. AIDS Clin. Res. 1 https://dx.doi.org/ Krishnamurti, T., London, E.D., Shoptaw, S., 2019. Contingency management
10.4172%2F2155-6113.S1-010. treatment for methamphetamine use disorder in South Africa. Drug Alcohol Rev. 39,
Corsi, K.F., Shoptaw, S., Alishahi, M., Booth, R.E., 2019. Interventions to reduce drug use 216–222. https://doi.org/10.1111/dar.13019.
among methamphetamine users at risk for HIV. Curr. HIV/AIDS Rep. 16, 29–36. Petry, N.M., 2010. Contingency management treatments: controversies and challenges.
https://doi.org/10.1007/s11904-019-00423-y. Addiction 105, 1507–1509. https://doi.org/10.1111/j.1360-0443.2009.02879.x.
Davis, D.R., Kurti, A.N., Skelly, J.M., Redner, R., White, T.J., Higgins, S.T., 2016. Petry, N.M., DePhilippis, D., Rash, C.J., Drapkin, M., McKay, J.R., 2014. Nationwide
A review of the literature on contingency management in the treatment of substance dissemination of contingency management: the veterans administration initiative.
use disorders, 2009–2014. Prev. Med. 92, 36–46. https://doi.org/10.1016/j. Amer. J. Addict. 23, 205–210. https://doi.org/10.1111/j.1521-0391.2014.12092.x.
ypmed.2016.08.008. Petry, N.M., Alessi, S.M., Rash, C.J., Barry, D., Carroll, K.M., 2018. A randomized trial of
DeFulio, A., Silverman, K., 2012. The use of incentives to reinforce medication contingency management reinforcing attendance at treatment: do duration and
adherence. Prev. Med. 55, S86–S94. https://doi.org/10.1016/j.ypmed.2012.04.017. timing of reinforcement matter? J. Consult. Clin. Psychol. 86, 799–809. https://doi.
Fletcher, J.B., Reback, C.J., 2013. Antisocial personality disorder predicts org/10.1037/ccp0000330.
methamphetamine treatment outcomes in homeless, substance-dependent men who Prendergast, M., Podus, D., Finney, J., Greenwell, L., Roll, J., 2006. Contingency
have sex with men. J. Subst. Abuse Treat. 45, 266–272. https://doi.org/10.1016/j. management for treatment of substance use disorders: a meta-analysis. Addiction
jsat.2013.03.002. 101, 1546–1560. https://doi.org/10.1111/j.1360-0443.2006.01581.x.
Fletcher, J.B., Dierst-Davies, R., Reback, C.J., 2014. Contingency management voucher Rawson, R., Huber, A., Brethen, P., Obert, J., Gulati, V., Shoptaw, S., Ling, W., 2000.
redemption as an indicator of delayed gratification. J. Subst. Abuse Treat. 47, 73–77. Methamphetamine and cocaine users: differences in characteristics and treatment
https://doi.org/10.1016/j.jsat.2014.03.003. retention. J. Psychoactive Drugs 32, 233–238. https://doi.org/10.1080/
Gómez, W., Olem, D., Andrews, R., Discepola, M.V., Ambrose, P., Dilworth, S.E., 02791072.2000.10400234.
Carrico, A.W., 2018. Optimizing contingency management with methamphetamine- Rawson, R.A., Gonzales, R., Ling, W., 2006. Methamphetamine abuse and dependence:
using men who have sex with men. Cogn. Behav. Pract. 25, 286–295. https://doi. an update. Dir. Psychiatry 26, 131–144.
org/10.1016/j.cbpra.2017.08.003. Reback, C.J., Shoptaw, S., 2014. Development of an evidence-based, gay-specific
Hedegaard, H., Miniño, A.M., Warner, M., 2020. Drug overdose deaths in the United cognitive behavioral therapy intervention for methamphetamine-abusing gay and
States, 1999-2018. NCHS Data Brief. bisexual men. Addict. Behav. 39, 1286–1291. https://doi.org/10.1016/j.
Higgins, S.T., Delaney, D.D., Budney, A.J., Bickel, W.K., Hughes, J.R., Foerg, F., addbeh.2011.11.029.
Fenwick, J.W., 1991. A behavioral approach to achieving initial cocaine abstinence. Reback, C.J., Peck, J.A., Dierst-Davies, R., Nuno, M., Kamien, J.B., Amass, L., 2010.
Am. J. Psychiatry 148, 1218–1224. https://doi.org/10.1176/ajp.148.9.1218. Contingency management among homeless, out-of-treatment men who have sex
Higgins, S.T., Budney, A.J., Bickel, W.K., Hughes, J.R., Foerg, F., Badger, G., 1993. with men. J. Subst. Abuse Treat. 39, 255–263. https://doi.org/10.1016/j.
Achieving cocaine abstinence with a behavioral approach. Am. J. Psychiatry 150, jsat.2010.06.007.
763–769. https://doi.org/10.1176/ajp.150.5.763. Reback, C.J., Peck, J.A., Fletcher, J.B., Nuno, M., Dierst-Davies, R., 2012. Lifetime
Higgins, S.T., Budney, A.J., Bickel, W.K., Foerg, F.E., Donham, R., Badger, G.J., 1994. substance use and HIV sexual risk behaviors predict treatment response to
Incentives improve outcome in outpatient behavioral treatment of cocaine contingency management among homeless, substance-dependent MSM.
dependence. Arch. Gen. Psychiatry 51, 568–576. https://doi.org/10.1001/ J. Psychoactive Drugs 44, 166–172. https://doi.org/10.1080/
archpsyc.1994.03950070060011. 02791072.2012.684633.
Higgins, S.T., Wong, C.J., Badger, G.J., Ogden, D.E.H., Dantona, R.L., 2000. Contingent Roll, J.M., Shoptaw, S., 2006. Contingency management: schedule effects. Psychiatry
reinforcement increases cocaine abstinence during outpatient treatment and 1 year Res. 144, 91–93. https://doi.org/10.1016/j.psychres.2005.12.003.
of follow-up. J. Consult. Clin. Psychol. 68, 64. https://doi.org/10.1037/0022- Roll, J.M., Huber, A., Sodano, R., Chudzynski, J.E., Moynier, E., Shoptaw, S., 2006a.
006X.68.1.64. A comparison of five reinforcement schedules for use in contingency management-
Kirby, K.C., Benishek, L.A., Dugosh, K.L., Kerwin, M.E., 2006. Substance abuse treatment based treatment of methamphetamine abuse. Psychol. Rec. 56, 67–81. https://doi.
providers’ beliefs and objections regarding contingency management: implications org/10.1007/BF03395538.
for dissemination. Drug Alcohol Depend. 85, 19–27. https://doi.org/10.1016/j. Roll, J.M., Petry, N.M., Stitzer, M.L., Brecht, M.L., Peirce, J.M., McCann, M.J., Blaine, J.,
drugalcdep.2006.03.010. MacDonald, M., DiMaria, J., Lucero, L., Kellogg, S., 2006b. Contingency
Krishnamurti, T., Murtaugh, K.L., Van Nunen, L., Davis, A.L., Ipser, J., Shoptaw, S., 2020. management for the treatment of methamphetamine use disorders. Am. J. Psychiatry
Spending money to make change: association of methamphetamine abstinence and 163, 1993–1999. https://doi.org/10.1176/ajp.2006.163.11.1993.
voucher spending among contingency management pilot participants in South Roll, J.M., Chudzynski, J., Cameron, J.M., Howell, D.N., McPherson, S., 2013. Duration
Africa. J. Subst. Abuse Treat. 112, 60–67. https://doi.org/10.1016/j. effects in contingency management treatment of methamphetamine disorders.
jsat.2020.01.014. Addict. Behav. 38, 2455–2462. https://doi.org/10.1016/j.addbeh.2013.03.018.
Kurti, A.N., Davis, D., Redner, R., Jarvis, B., Zvorsky, I., Kieth, D.R., Bolivar, H., White, T. Sayegh, C.S., Huey Jr., S.J., Zara, E.J., Jhaveri, K., 2017. Follow-up treatment effects of
J., Rippberger, P., Markeish, C., Atwood, G., Higgins, S.T., 2015. A review of the contingency management and motivational interviewing on substance use: a meta-
literature on remote monitoring technology in incentive-based interventions for analysis. Psychol. Addict. Behav. 31, 403–414. https://doi.org/10.1037/
health-related behavior change. Transl. Issues Psychol. Sci. 2, 128–152. https://doi. adb0000277.
org/10.1037/tps0000067. Shoptaw, S., Reback, C.J., Peck, J.A., Yang, X., Rotheram-Fuller, E., Larkins, S.,
Lake, M.T., Shoptaw, S., Ipser, J.C., Takada, S., van Nunen, L.J., Lipinska, G., Stein, D.J., Viniegas, R.C., Freese, T.E., Hucks-Ortiz, C., 2005. Behavioral treatment approaches
London, E.D., 2020. Decision-making by patients with methamphetamine use for methamphetamine dependence and HIV-related sexual risk behaviors among
disorder receiving contingency management treatment: magnitude and frequency urban gay and bisexual men. Drug Alcohol Depend. 78, 125–134. https://doi.org/
effects. Front. Psychiatry 11, 22. https://doi.org/10.3389/fpsyt.2020.00022. 10.1016/j.drugalcdep.2004.10.004.
Landovitz, R.J., Fletcher, J.B., Inzhakova, G., Lake, J.E., Shoptaw, S., Reback, C.J., 2012. Shoptaw, S., Huber, A., Peck, J., Yang, X., Liu, J., Dang, J., Roll, J., Shapiro, B.,
A novel combination HIV prevention strategy: post-exposure prophylaxis with Rotherham-Fuller, E., Ling, W., 2006a. Randomized, placebo-controlled trial of
contingency management for substance abuse treatment among methamphetamine- sertraline and contingency management for the treatment of methamphetamine
using men who have sex with men. AIDS Patient Care STDS 26, 320–328. https:// dependence. Drug Alcohol Depend. 85, 12–18. https://doi.org/10.1016/j.
doi.org/10.1089/apc.2011.0432. drugalcdep.2006.03.005.
Lee, N.K., Rawson, R.A., 2008. A systematic review of cognitive and behavioural Shoptaw, S., Klausner, J.D., Reback, C.J., Tierney, S., Stansell, J., Hare, C.B., Gibson, S.,
therapies for methamphetamine dependence. Drug Alcohol Rev. 27, 309–317. Siever, M., King, W.D., Kao, U., Dang, J., 2006b. A public health response to the
https://doi.org/10.1080/09595230801919494. methamphetamine epidemic: the implementation of contingency management to
Ling Murtaugh, K., Krishnamurti, T., Davis, A.L., Reback, C.J., Shoptaw, S., 2013. Spend treat methamphetamine dependence. BMC Public Health 6, 214. https://doi.org/
today, clean tomorrow: predicting methamphetamine abstinence in a randomized 10.1186/1471-2458-6-214.
controlled trial. Health Psychol. 32, 958. https://doi.org/10.1037/a0032922.

12
H.D. Brown and A. DeFulio Drug and Alcohol Dependence 216 (2020) 108307

Shoptaw, S., Reback, C.J., Larkins, S., Wang, P.-C., Rotherham-Fuller, E., Dang, J., et al., Substance Abuse and Mental Health Services Administration, 2019. Key Substance Use
2008. Outcomes using two tailored behavioral treatments for substance abuse in and Mental Health Indicators in the United States: Results from the 2018 National
urban gay and bisexual men. J. Subst. Abuse Treat. 35, 285–293. https://doi.org/ Survey on Drug Use and Health [Annual Report]. Retrieved from:. U.S. Department
10.1016/j.jsat.2007.11.004. of Health and Human Services https://www.samhsa.gov/data/report/2018-nsduh
Silverman, K., Wong, C.J., Umbricht-Schneiter, A., Montoya, I.D., Schuster, C.R., -annual-national-report.
Preston, K.L., 1998. Broad beneficial effects of cocaine abstinence reinforcement Behavioral pharmacology of human drug dependence. In: Thompson, T., Johanson, C.E.
among methadone patients. J. Consult. Clin. Psychol. 66, 811–824. https://doi.org/ (Eds.), 1981. NIDA Research Monograph 37. U.S. Government Printing Office,
10.1037/0022-006X.66.5.811. Washington, D.C. https://archives.drugabuse.gov/sites/default/files/monograph37.
Silverman, K., DeFulio, A., Sigurdsson, S.O., 2012. Maintenance of reinforcement to pdf.
address the chronic nature of drug addiction. Prev. Med. 55, S46–S53. https://doi. Vulimiri, M., Bleser, W.K., Saunders, R.S., Madanay, F., Moseley, C., McGuire, H.F.,
org/10.1016/j.ypmed.2012.03.013. Ubel, P.A., McKethan, A., McClellan, M., Wong, C.A., 2019. Engaging beneficiaries
Silverman, K., DeFulio, A., Leoutsakos, J.M., 2013. The potential value of a negative in Medicaid programs that incentivize health-promoting behaviors. Health Aff. 38,
finding: an illustrative example. JAMA Psychiatry 70, 571–572. https://doi.org/ 431–439. https://doi.org/10.1377/hlthaff.2018.05427.
10.1001/jamapsychiatry.2013.1116. Zhang, S.X., Shoptaw, S., Reback, C.J., Yadav, K., Nyamathi, A.M., 2018. Cost-effective
Strona, F.V., McCright, J., Hjord, H., Ahrens, K., Tierney, S., Shoptaw, S., Klausner, J.D., way to reduce stimulant-abuse among gay/bisexual men and transgender women: a
2006. The acceptability and feasibility of the Positive Reinforcement Opportunity randomized clinical trial with a cost comparison. Public Health 154, 151–160.
Project, a community-based contingency management methamphetamine treatment https://doi.org/10.1016/j.puhe.2017.10.02.
program for gay and bisexual men in San Francisco. J. Psychoactive Drugs 38,
377–383. https://doi.org/10.1080/02791072.2006.10400601.

13