Just-In-Time Adaptive Interventions for Addictive Behaviors

BACKGROUND: Our research group has made extensive use of ecological momentary assessment
(EMA) to understand people s daily experiences with opioid craving, use, and lapse. We have
also administered clinic-based psychotherapies such as cognitive-behavioral therapy (CBT), in
which clients learn to avoid, escape, resolve, or reframe problems (such as stress) that can
trigger lapses to drug use. CBT contrasts with ACT (Acceptance and Commitment Therapy), a
mindfulness-based approach in which difficult thoughts and feelings are viewed as necessary
and potentially valuable components of a full life, to be experienced observantly rather than
resolved or reframed. CBT and ACT can and do coexist in a single treatment plan, but we know
of no systematic attempt to reconcile their differences. We have also not tried to administer
either of them on a mobile device.

OBJECTIVE: To test a just-in-time adaptive intervention (JITAI)-a treatment given when and
where it is needed. Our JITAI will be delivered via smartphone app and will combine elements
of two widely used treatments for addiction: CBT and ACT. Our goal is not to bring another
branded app onto the market, but rather to clarify when and for whom the generic components
of such apps are effective or not. This will include determining when CBT is more helpful
than ACT and vice versa.

PARTICIPANT POPULATION: Outpatient adults who are physically dependent on opioids -up to 175
enrolled (25 for a formative-interview phase, 150 for a trial) for a target of 105 evaluable
(20 interviewees, 85 trial participants). Target enrollment will include 40% women and 60%
minorities (mostly African-American). In the trial, some participants will receive
buprenorphine in our clinic, and others will be receiving buprenorphine or methadone
elsewhere; this is a procedural matter, not a component of the experimental design.

EXPERIMENTAL DESIGN: After a formative-interview phase, the study will be run as a
microrandomized trial that will also include a conventionally randomized between-groups
clinical-trial component. In microrandomization, interventions are randomized at the
momentary level within person; the effect is measured proximally (e.g., 20 minutes later).
This is a powerful way to assess the effects of different interventions administered in the
field and to examine strategy-situation fit, i.e., whether interventions are differentially
effective under specific momentary circumstances. We are powering our study mostly to detect
(1) any effect of CBT or ACT versus control moments with no intervention given, and (2)
preferential advantages of CBT over ACT, and vice versa, as a function of the participant s
ability to control (change, escape) a given situation. The between-groups aspect of the
design (JITAI group versus EMA-only control group) is needed to demonstrate an effect of our
JITAI on traditional, distal measures of outcome, such as reductions in opioid use.

METHODS: In the formative-interview phase, we will conduct interviews with people who use
opioids and express interest in using a mobile treatment app. We will show interviewees item
lists, onscreen mockups, and/or functional demos, and we will ask interviewees to comment on
the app's likely usefulness, its likely pitfalls, and how we could improve it.

In the clinical trial, participants will be randomized to one of two groups (JITAI vs.
EMA-only control). During weeks 1-2, all participants will have baseline assessments of
coping styles and personality, and all JITAI participants will be shown a video introducing
basic concepts of CBT and ACT. For weeks 3-10 (8 weeks), participants will carry smartphones
for EMA with or without JITAI. During week 11, participants will be readministered some of
the assessments from baseline. Participants receiving buprenorphine from us will then be
offered a dose taper or encouraged to transfer to continued treatment elsewhere. All
participants will come to our clinic thrice weekly for urine testing throughout
participation.

PRIMARY OUTCOME MEASURES: (1) Proximal effects of CBT and ACT messages in the JITAI group:
decreases over 20-minute intervals in craving and negative mood, with increases in
self-efficacy; (2) strategy-situation fit in the JITAI group; (3) group differences in distal
effects of treatment (week 11 versus week 2) in terms of self-efficacy and coping
flexibility.

SECONDARY OUTCOME MEASURES: (1) Trait predictors of differential responses to CBT and ACT, in
the JITAI group; (2) group differences in frequency of opioid-positive urine over time; (3)
time courses of responsiveness to ACT vs. CBT, in the JITAI group; (4) whether the
intervention types that benefit participants most when pushed by the app are the same ones
participants choose when subsequently given the opportunity to pull interventions.

- INCLUSION CRITERIA:

Phase 1: Formative interviews.

The enrollment ceiling is 25 outpatients (to collect evaluable data from 20) who meet these
criteria: (1) Age 18-75; (2) physical dependence on opioids (by self-report); (3) interest
in receiving the types of treatment about which we will be conducting interviews.

Phase 2: Clinical trial with microrandomization.

The enrollment ceiling is 150 outpatients (to collect evaluable data from 85, of whom 50
will be randomized to JITAI, and 35 to EMA control). Treatment may be provided by us in the
form of office-based buprenorphine treatment (OBOT) or may be provided elsewhere (Treatment
Elsewhere, TE). Participants must meet these criteria:

OBOT participants: (1) Age 18-75; (2) physical dependence on opioids (by positive urine
and/or frank opioid withdrawal); (3) interest in receiving the types of treatment we are
testing.

Treatment Elsewhere (TE) participants: (1) Age 18-75; (2) receiving methadone or
buprenorphine treatment for opioid dependence from a qualified provider in the community;
(3) interest in receiving the types of treatment we are testing.

EXCLUSION CRITERIA:

Phase 1: Formative interviews. (1) cognitive impairment severe enough to preclude informed
consent or valid interview responses.

Phase 2: Clinical trial with microrandomization.

OBOT participants: (1) History of any DSM-V psychotic disorder; history of bipolar
disorder; current Major Depressive Disorder; (2) unresolved symptoms of PTSD that, in the
investigators view, would make it risky for the participant to undertake mindfulness
exercises (e.g., observing all one s current negative thoughts and emotions) in an
unsupervised setting; (3) current dependence on alcohol or sedative-hypnotic, e.g.
benzodiazepine (by DSM-V criteria); (4) cognitive impairment severe enough to preclude
informed consent or valid self-report; (5) Any condition that interferes with urine
collection; (6) medical illness (e.g., cirrhosis, nephritic syndrome, thyroid disease,
ischemic heart disease, epilepsy, adrenal insufficiency, etc.) or medications that, in the
view of the investigators, would compromise participation in research.

Treatment Elsewhere (TE) participants: (1) History of any DSM-V psychotic disorder; history
of bipolar disorder; current Major Depressive Disorder; (2) unresolved symptoms of PTSD
that, in the investigators view, would make it risky for the participant to undertake
mindfulness exercises (e.g., observing all one s current negative thoughts and emotions) in
an unsupervised setting; (4) current dependence on alcohol or sedative-hypnotic, e.g.
benzodiazepine (by DSM-V criteria); (5) cognitive impairment severe enough to preclude
informed consent or valid self-report; (6) Any condition that interferes with urine
collection; (6) medical illness (e.g., cirrhosis, nephritic syndrome, thyroid disease,
ischemic heart disease, epilepsy, adrenal insufficiency, etc.) or medications that, in the
view of the investigators, would compromise participation in research.