Computer-Facilitated Screening and Brief Intervention for Alcohol Use Risk in Adolescent Patients of Pediatric Primary Care Offices: Protocol for a Cluster Randomized Controlled Trial

Background Alcohol and other substance use disorders usually begin with substance use in adolescence. Pediatric primary care offices, where most adolescents receive health care, are a promising venue for early identification of substance use and for brief intervention to prevent associated problems and the development of substance use disorder. Objective This study tests the effects of a computer-facilitated screening and brief intervention (cSBI) system (the CRAFFT [Car, Relax, Alone, Forget, Family/Friends, Trouble] Interactive System [CRAFFT-IS]) on heavy episodic drinking, riding with a driver who is substance impaired, or driving while substance impaired among adolescents aged 14 to 17 years presenting for a well visit at pediatric primary care practices. Methods We are conducting a cluster randomized controlled trial of the CRAFFT-IS versus usual care and recruiting up to 40 primary care clinicians at up to 20 pediatric primary care practices within the American Academy of Pediatrics (AAP) Pediatric Research in Office Settings network. Clinicians are randomized 1:1 within each practice to implement the CRAFFT-IS or usual care with a target sample size of 1300 adolescent patients aged 14 to 17 years. At study start, intervention clinicians complete web-based modules, trainer-led live sessions, and mock sessions to establish baseline competency with intervention counseling. Adolescents receive mailed recruitment materials that invite adolescents to complete an eligibility survey. Eligible and interested adolescents provide informed assent (parental permission requirement has been waived). Before their visit, enrolled adolescents seeing intervention clinicians complete a self-administered web-based CRAFFT screening questionnaire and view brief psychoeducational content illustrating substance use–associated health risks. During the visit, intervention clinicians access a computerized summary of the patient’s screening results and a tailored counseling script to deliver a motivational interviewing–based brief intervention. All participants complete previsit, postvisit, and 12-month follow-up study assessments. Primary outcomes include past 90-day heavy episodic drinking and riding with a driver who is substance impaired at 3-, 6-, 9-, and 12-month follow-ups. Multiple logistic regression modeling with generalized estimating equations and mixed effects modeling will be used in outcomes analyses. Exploratory aims include examining other substance use outcomes (eg, cannabis and nicotine vaping), potential mediators of intervention effect (eg, self-efficacy not to drink), and effect moderation by baseline risk level and sociodemographic characteristics. Results The AAP Institutional Review Board approved this study. The first practice and clinicians were enrolled in August 2022; as of July 2023, a total of 6 practices (23 clinicians) had enrolled. Recruitment is expected to continue until late 2024 or early 2025. Data collection will be completed in 2025 or 2026. Conclusions Findings from this study will inform the promotion of high-quality screening and brief intervention efforts in pediatric primary care with the aim of reducing alcohol-related morbidity and mortality during adolescence and beyond. Trial Registration ClinicalTrials.gov NCT04450966; https://www.clinicaltrials.gov/study/NCT04450966 International Registered Report Identifier (IRRID) DERR1-10.2196/55039


PUBLIC HEALTH RELEVANCE
Our goal is to conduct a large multi-site randomized controlled trial (RCT) of a promising computerfacilitated Screening and clinician Brief Intervention (cSBI) system designed for delivery by pediatric primary care clinicians and aimed at reducing unhealthy alcohol use and related riding/driving safety risk among adolescent patients.Our setting will be the American Academy of Pediatrics' (AAP) Pediatric Research in Office Settings (PROS) national primary care research network, with U.S. primary care practices having participated in recent studies.This trial addresses the evidence gap identified in the latest U.S. Preventive Services Task Force review of alcohol screening and brief counseling interventions among adolescents, and, if shown effective, the cSBI system could be widely disseminated via AAP's existing education, teaching, and advocacy platforms to its 67,000 pediatrician members, thereby greatly increasing the potential for population-level impact of alcohol screening and brief intervention for U.S. adolescents.

Overall Impact:
This is a resubmission of a multisite study to evaluate computerized SBI to standard care in a large and diverse sample of adolescents (aged 14-17) in a range of outpatient pediatric primary care settings associated with the AAP research network.In the previous reviews, strengths included an outstanding team, access to a nationwide network which would provide access a large sample size, and a strong environment.Inclusion of a TAU control and a one-year follow-up were also strengths, as would be the AAP network for dissemination.
Weaknesses identified in the previous reviews were multiple and significant.These included lack of innovation regarding E-SBI, the multi-component approach to be evaluated with several untested components included, lack of rationale for cannabis use as a secondary aim, questions about incremental gains to knowledge for a fairly well-studied area, and whether the modest effects of the previous study by this team justified a large, multisite trial.Reviewers pointed out that evaluation of long-term harms would be needed generate data that would affect policy.The texting component was not well developed or integrated.There were also major concerns about plans to deal with contamination.Practical issues noted included the need for more attention to fidelity ratings and training of staff in multiple sites.
The resubmission was responsive to many of the issues raised in the prior review.The start-up period was shortened, the untested aspects of the intervention were dropped, randomization would be via clinician rather than patient to reduce contamination.The investigators clarified that AAP policy is that network sites can only be selected after a project is funded and IRB approval is in place; the PROS network is established and has its own single IRB, another plus.The waiver rather than parental consent appears justified given the low risk and the nature of this study.The project, following advice from reviewers, is now better positioned to address USPSTF issues related to SBI in adolescents, specifically around drinking related risks and harms in this population.The intervention itself is simpler and less of a "kitchen sink" approach.
However, the issues of innovation and incremental value to the data to be gained were not fully addressed.The large study evaluating the cSBI intervention is still under review; the data presented in the application was not particularly compelling and there was no strategy given for strengthening the approach.Fidelity evaluation remained weak.Outcomes would be entirely based on self-report.No 1 R01 AA027253-01A1 4 AA-3 HARRIS, S cost-effectiveness aspect was included.Hence, the approach still appears somewhat incremental, only adding a few new features, rather than a 'definitive' study of cSBI for adolescents.

Strengths
• As before, using pediatric primary care visits as a conduit to prevent or reduce risky alcohol in adolescents is of high public health significance.
• As before, universal screening regardless of drinking status adds impact; however • As before, delivering via computer should facilitate dissemination if the approach demonstrates significant effects on drinking, health, and risk outcomes.

Weaknesses
• As noted above, the previous, similar, large multisite study of the approach remains under review.The investigators did not make a strong case regarding important novel aspects of this work.

Strengths
• As before, the investigative team is strong and clearly capable of carrying out the proposed work.
• Dr. Harris is a productive young investigator with almost 80 peer reviewed publications, most in areas related to the proposed work.

Strengths
• Evaluation of SBI is not particularly innovative and there are several studies in this area.The cSBI approach to be evaluated here has been fairly well studied before, with modest effects.

Strengths
• As before, the practice care network, cSBI approach, follow-up, sample size, and other aspects noted are all strengths that still pertain.The sample (1268 adolescents aged 14-17 in diverse areas across the US) ample, and large and diverse enough for generalizability.
• Retention rates in prior similar studies by this group have been good (approaching 80%).
• The cSBI intervention feedback and brief interventions are somewhat tailored to the adolescents' reported drinking level.
1 R01 AA027253-01A1 5 AA-3 HARRIS, S • Clinicians will be randomized within practice; again, however, since SBIRT activities have been around for many years and are widely disseminated and recommended by AAP, its likely that many of the clinicians not randomized to cSBI will be doing a reasonably good job of screening and intervening with adolescents at risk, reducing likelihood of a larger effect.
• Training of the clinicians assigned to the cSBI approach is reasonable without being overly burdensome.

Weaknesses
• The time to drinking data presented in Figure 2 did not indicate a significant effect, leading to questions about the strength and promise of the cSBI intervention.Again, since the main report remains under review and is not available to reviewers, the present studies may be premature and justification was not compelling.
• The investigators appear to be moving away from universal screening by including adolescents with past-year drinking or riding in a car with an impaired driver, which may reduce generalizability.
• The fidelity rating plan is improved but remains fairly passive and not consistent with standard methods of assessing fidelity to behavioral interventions.The rating form to be used was not specified and it was somewhat surprising the investigators had not developed this in prior research.
• Outcomes will be entirely based on self-report rather than medical records.The rationale for not collecting biological samples was not compelling and a major weakness that could undermine credibility of findings and their persuasiveness to the USPSTF.
• The statistical plan does not consider the possibility of interventionist (clinician) effects.
• It was not clear why a cost-effectiveness analysis was not involved, again reducing likely impact of results.

Strengths
• As before, the environment is strong, no concerns.

Strengths
• Seems reasonable, if ambitious.Long lead in period was reduced.

Protections for Human Subjects:
Acceptable Risks and/or Adequate Protections • As before, it is clear that the team has experience to conduct this trial safely with multiple subject protections in place.Using the single IRB associated with the AAP network is a plus.This revised proposal from an NIH NEW INVESTIGATOR is for a cluster-randomized efficacy trial of pediatricians, testing a computerized alcohol screening for alcohol use, followed by computerized brief education (cSBI) followed by pediatrician brief advice.The proposal explicitly targets the USPSTF call for more research on alcohol SBI in adolescents.The trial (in 10 PROS practices) evaluates the intervention in an estimated 1268 "at-risk" patients between ages 14-17 seeing 30 pediatricians who enroll and who themselves consent.Patient eligibility is determined based on annual well visit, screening positive for past 12-month alcohol use, or riding risk, and having cell phone.The trial is generally rigorously designed, and the revision is responsive to many prior critiques.Issues from the prior reviews that were addressed include: a conceptual model, omission of add-on texting interventions, how MI is incorporated, explanation of training of the pediatricians and rigorous monitoring of fidelity.The revision also explains that the prior "trial" was actually a pre-post study with major flaws.A waiver of consent from patients appears to be feasible.
Major limitations of the current proposal, however, persist.Despite saying that the USPSTF is a major target audience of this trial, the trial does not address why the PIs prior trial was not included in the recent USPSTF review.It appears it was deemed "prevention" because the sample (apparently all pediatric patients) was not of interest to USPSTF.The current proposal focuses on patients who have used any alcohol in the past year, but the outcome is heavy episodic drinking (any).It is unclear that this sample will be acceptable to USPSTF and that is not addressed.No data in the Approach section indicates this is an adequate screen for unhealthy alcohol use -the NIAAA screen is noted to be a valid screen for any use past year (page 116).This could still be perceived as prevention by USPSTF, especially since the outcome is not any use.While trials of BIs are needed with adolescents with unhealthy alcohol use, this proposal does not clearly demonstrate that it fills that gap.• USPSTF has highlighted the need for high quality trials of youth alcohol SBI in primary care • Evidence suggests that SBI might be efficacious.
• Computerized screening has advantages in terms of privacy and workflow • PI's group previously conducted a pre-post "trial" with 18 months of pre, no apparent washout (to avoid intervention after conversion of the clinic), exclusion of patients from the post period (resulting in differences in the 2 samples in terms of drinking), and evaluating drinking at 3 and 12 months (not HED).Certainly, that is not a rigorous design, suggesting that further study is warranted.
• Premise: Need for high quality studies is strong.
• References to USPSTF calls for more research (KQ#4 and KQ#5) are helpful.and P30s.Dr. Fiks will lead the PROS team contributing to design, implementation, analysis and manuscript preparation.

Weaknesses
• Dr. Knight (developer of the CRAFT and founder of CeASAR) retired 12/31/2017 but will be a consultant (25-100 hours a year) to support Dr. Harris.
• Lon Sherritt MPH (1999) has designed and conducted analyses as well as web platforms and NLP algorithms for CeASAR since its inception, • Dr. Shone (PhD, MSW) is the Director of Primary Care Research for the AAP and will oversee the trial for AAP.

Weaknesses
• Minor: While several investigators have experience with clinical trials, there is no clear trialist on the team.While the generally strong protocol suggests that there might be adequate expertise, this reviewer would like to see a PhD-level statistician with experience in pediatric trials.

Strengths
• Use of AAP PROS for alcohol research is novel

Weaknesses
• This is incremental research that does not address the challenge of unhealthy alcohol use in adolescents.Although it is sorely needed.All elements have been studied in prior studies by the PIs team-albeit with inadequate rigor.

Strengths
• Strong intervention.Combining technology for screening and psycho-ed feedback, with primary care clinicians' advice, seems like a strength.• The trial is rigorously designed and described including recruitment, the intervention, rationale for waiver of parental consent, • While USPSTF has relied on consumption outcomes for adults, measures of harms are encouraged, and have been added.

Weaknesses
• Sample.Including patients who only used 1 or more times in the past year (irrespective of age) will make it hard to show benefit (see figure on consumption and outcomes in the USPSTF 2018).No evidence is presented that this is a validated screen for unhealthy alcohol use.That could also make the USPSTF reviewers consider it primary prevention (which appears to make them omit the PIs study in Pediatrics in 2012).Screening for unhealthy alcohol use and/or AUDs is strongly advised if the goal is to decrease heavy drinking at 12 months follow-up.Also, 1 R01 AA027253-01A1 9 AA-3 HARRIS, S no preliminary data are recalled by this reviewer to demonstrate high rates of heavy episodic drinking during follow-up among those with any past 12-month use.
• Preliminary studies of the intervention (cSBI) were described as an RCT that was really a prepost evaluation.It did not find cSBI efficacious (pre-post).That study had a number of limitations: patients in phase 1 (UC) were removed from phase 2 (intervention) and there were significant differences between the groups, as well as lack of a washout period between the 18 months.A quick read did not identify any randomization (as in a stepped wedge design).
• Measures.Measuring DSM-5 AUD symptoms at baseline would be critically important.
Outcomes may vary depending on AUD v not.
• This reviewer would have liked to have a clear statement that the trial was addressing and meeting USPSTF criteria for the recent 2018 review on alcohol interventions.It appears it would not have.The randomized clinical trial proposed in this revised application builds on prior work that showed a multi-faceted computerized screening tool paired with direct face-to-face counseling delivered in pediatric primary care offices reduced heavy episodic drinking among adolescents who reported pastyear alcohol use at baseline.This proposal aims to replicate the initial study in a larger sample that is geographically and socioeconomically diverse.Identifying promising new ways to reach adolescents, an especially vulnerable population in the context of alcohol use and misuse, could have significant clinical and public health implications.The prior submission was considered meritorious based on compelling preliminary data that supported the feasibility and efficacy of the approach, the aim to leverage pediatric practices as a natural point of engagement for youth, the careful selection of measures, and the overall study design.Prior concerns were related to the unconvincing justification for a second large-scale randomized clinical trial that would essentially replicate recent work by the investigative team.In addition, there were concerns that this new trial would still prove insufficient to move the US Preventative Task Force to recommend adolescent screening and brief intervention (a stated goal of the applicants).Additional questions were raised regarding the multiple untested additions to the initial brief intervention (especially the texting protocol), clinician training and fidelity monitoring, the scope and feasibility of the proposed work given that specific clinic sites were not identified and the staffing plan seemed logistically challenging, and insufficient attention to health-risks associated with adolescent drinking.Many of these concerns were carefully considered and sufficiently addressed in the revised application.Concerns remain, however, about the justification for a second large-scale trial.In addition, it is unclear how preliminary findings relate to the proposed work given that participants in the initial trial were required to report past 12-month drinking at baseline.The proposed study will enroll nondrinkers along with drinkers and inadequate attention is given to how these two distinct groups of youth will be handled in analyses.In addition, inasmuch as the proposed computerized screening and in-person brief intervention requires considerable time, there is question about how well this approach -even if efficacious -will translate to typical pediatrician practices.Indeed, the proposed approach requires clinicians to complete several hours of intensive training, establish baseline competency, and engage in ongoing adherence monitoring and expert feedback.On the whole, however, the investigators were responsive to prior critiques, the proposed work could have substantive impact clinical care for adolescents, and there is no question the investigative team could successfully execute the proposed research.

•
Data and Safety Monitoring Plan (Applicable for Clinical Trials Only): Sex/Gender: Distribution justified scientifically • Race/Ethnicity: Distribution justified scientifically • For NIH-Defined Phase III trials, Plans for valid design and analysis: Not applicable • Inclusion/Exclusion Based on Age: Distribution justified scientifically Resource Sharing Plans: Acceptable Budget and Period of Support:

•
Unhealthy alcohol use in youth is associated with marked adverse impacts on like and heath

••
Outcomes: Objective outcome measures are critically important.Relying on patient self-report allows social desirability bias to impact the trial.Perhaps change in grades could be added as a secondary outcome? 5. Environment: Strengths • CeASAR has a long track record of research in this area.• The APP PROS has a track record for office-based trials.Obtaining central IRB (AAPs IRB) approval would help reviewers know that a waiver of parental consent is possible Data and Safety Monitoring Plan (Applicable for Clinical Trials Only): Acceptable Inclusion Plans: • Sex/Gender: Distribution justified scientifically • Race/Ethnicity: Distribution justified scientifically • For NIH-Defined Phase III trials, Plans for valid design and analysis: Not applicable • Inclusion/Exclusion Based on Age: Distribution justified scientifically 1

•
Literature and preliminary studies review unclear and deficient.This reviewer had to review the Pediatrics 2012 manuscript to understand the above, indicating that the proposal did not still delineate characteristics of the prior studies and the proposed study and the 2 trials reviewed by USPSTF in 2018.A table of sample, intervention, outcomes and follow-up would have helped understand what USPSTF requires.
• Specifically, the limitations in the extant literature would ideally have been much more clearly summarized, including the USPSTF review (2 studies) and trials since.Only 2 trials met quality criteria for the USPSTF review 2018 which admittedly came out several months before this proposal was submitted.It would have been helpful to clearly state how the proposed trial differed and how it was the same as those reviewed and those excluded.(Mason M, Light J, Campbell L, et al.Peer Network Counseling with Urban Adolescents: A Randomized Controlled Trial with Moderate Substance Users.J Subst Abuse Treat.2015;58:16-24; Haug S, Paz Castro R, Kowatsch T, et al.Efficacy of a Web-and Text Messaging-Based Intervention to Reduce Problem Drinking in Adolescents: Results of a Cluster-Randomized Controlled Trial.J Consult Clin Psychol.2016.)• Primary

Prevention Trials not of interest to USPSTF.
The prior trial by the PI (Pediatrics 2012) was not included in the USPSTF review (reason "prevention").It would have been helpful if the proposal delineated whether the proposed trial was designed in a manner where it would have overcome limitations of the prior study and been eligible for inclusion in the 2018 review.Any alcohol use past year seems like primary prevention to this reviewer given that the outcome is heavy episodic drinking.Page 12 clearly cites USPSTF interest in interventions that "reduce unhealthy alcohol use".The decision to define any past year use as unhealthy alcohol use might have been considered acceptable if the trial used that as an outcome.However, this reviewer does not expect most experts would define unhealthy alcohol use as any past year use (or at least it would need to be clear whether or not that has not been the gold standard for unhealthy alcohol use in validation studies of alcohol screens for unhealthy alcohol use in adolescents).She is a new NIH investigator, having been made a MPI after submission for the prior study.•Dr. Burke (Northeastern U Clinical Professor since 2014 and PhD Program Director and MINT) brings MI expertise and will provide clinician trainings.She has done the trainings for a grant with Sharon Levy and Eli Weitzman, as well as the Moment study (Shrier PI) Strengths• Dr. Harris (PhD Johns Hopkins 1996), Assoc.Professor Pediatrics Harvard Medical School (2017-), and Director of CeASAR, Boston Children's Hospital since 2018 when Dr. Knight