The Opioid Analgesic Reduction Study (OARS): A Comparison of Opioid vs. Non-Opioid Combination Analgesics for Management of Post-Surgical Pain Study Protocol for a Double-Blind Randomized Multi-Center Clinical Trial

Background: Everyday people die unnecessarily from opioid overdose-related addiction. Dentists are among the leading prescribers of opioid analgesics. Opioid seeking behaviors have been linked to receipt of initial opioid prescriptions following the common dental procedure of third molar extraction. With each opioid prescription a patient’s risk for opioid misuse or abuse increases. With an estimated 56 million tablets of 5 mg hydrocodone annually prescribed after third molar extractions in the United States, 3.5 million young adults may be unnecessarily exposed to opioids by dentists who are inadvertently increasing their patient’s risk for addiction. Methods: A double blind, stratied randomized, multi-center clinical trial has been designed to evaluate whether a combination of over-the-counter non-opioid containing analgesics is not inferior to the most prescribed opioid analgesic. The impacted 3rd molar extraction model is being used due to the predictable severity of the post-operative pain and generalizability of results. Within each site/clinic and gender type (male/female), patients are randomized to receive either OPIOID (hydrocodone/acetaminophen 5/300 mg) or NON-OPIOID (ibuprofen/acetaminophen 400/500 mg). Outcome data include pain levels, adverse events, overall patient satisfaction, ability to sleep, and ability to perform daily functions. To develop clinical guidelines and a clinical decision making tool, pain management, extraction diculty and number of tablets taken is being collected enabling an experimental decision making tool to be developed. Discussion: The proposed methods address the short comings of other analgesic studies. Although prior studies have tested short-term effects of single doses of pain medications, patients and their dentists are interested in managing pain for the entire postoperative period, not just the rst 12 hours. After surgery, patients expect to be able to perform normal daily functions without feeling nauseous or dizzy and they desire a restful sleep at night. Parents of young people are concerned with the risks of opioid use and misuse, either related to treatments received or to subsequent use of leftover pills. Upon successful completion of this clinical trial, dentists, patients, and their families will be better able to make informed decisions regarding post-operative pain management. patient or others who may have access to the medications increases. Recent studies have shown that exposing young adults to an opioid analgesic increases their risk of future opioid use, with as much as 37% of non-medical opioid use by high school seniors originating from leftover opioid prescriptions in their or their friend’s or families’ households. (44) If an alternative analgesic is found to be non-inferior to the most commonly prescribed opioid for impacted third molar extractions, the overall number of opioid prescriptions would be reduced. In turn, with decreased opioid substrate remaining in the community, the number of young adults at risk for opioid abuse and addiction via drug diversion has the potential to be signicantly reduced. peptic ulcer; renal disease (excluding kidney stones); hepatic disease; cardiovascular disease (myocardial infarction or stroke within the past 6 months); bleeding disorder; respiratory depression; prior respiratory effect of an opioid or other anesthetic drugs required respiratory support postoperatively; active or untreated asthma; allergic reaction to ibuprofen, acetaminophen, hydrocodone, and/or anesthesia, consumption of alcoholic of


Introduction
The goal of this study is to provide health care professionals, including dentists, with the best possible evidence for clinical decision making when selecting analgesics for acute post-surgical pain management. A double-blind, strati ed randomized clinical trial is being conducted to test the hypothesis that a combination of over-the-counter non-opioid containing analgesics is at least as, if not more, effective (non-inferior) than the most commonly prescribed opioid analgesic. The impacted third molar extraction model is used due to the predictable severity of post-operative pain and generalizability of results, as well as the fact that dentists prescribe approximately one-third of all opioid prescriptions for adolescents in the United States.

Background and rationale {6a}
Description of Health Problem: Opioid-related deaths are rising alarmingly in the United States. (1) Opioid overdose deaths have more than quadrupled since 1999 (2) accounting for over 49,860 deaths in 2019. (3) Unprecedented increases in opioid overdose deaths have occurred during the COVID-19 pandemic with a 57.7% increase between May 2019 and May 2020. (4) Between 1999 -2019, nearly 247,000 Americans have died due to overdoses involving prescription opioids. (5) Many patients do not properly secure or dispose of their unused opioid medications, leaving them accessible to others who may initiate or feed an addiction .(6- 7) Millions of tablets of 5 mg hydrocodone-containing combinations are prescribed after third molar extractions each year in the United States. (8) High school students who receive an opioid prescription are 33% more likely than those who do not receive an opioid prescription to eventually misuse opioids (9) contributing to an upsurge in deaths among 18 to 25-year olds. (10) Dentists are among the leading prescribers of opioid analgesics, (11) accounting for 12% of all prescriptions for immediate-release opioids. (12) They rank fourth among medical specialties for their opioid prescribing rates, writing 18.5 million opioid prescriptions per year (13) . Over 76% of these opioid prescriptions have been for hydrocodone combination products. Adolescents and young adults receive more than 11% of dentist-prescribed opioids. (14) This nding of opioid-prescribing prevalence for adolescents is consistent with other studies and the assessment of acute opioid prescriptions for youth. (11,15) Dentists write about 31% of opioid prescriptions for young patients aged 10 to 19 (8) with about 61% of 14-to 17-year-olds receiving opioid prescriptions following third-molar extractions. Exclusion Criteria: Participants who self-report the following history are excluded from study participation: history of gastrointestinal bleeding and/or peptic ulcer; renal disease (excluding kidney stones); hepatic disease; cardiovascular disease (myocardial infarction or stroke within the past 6 months); bleeding disorder; respiratory depression; prior respiratory effect of an opioid or other anesthetic drugs that required respiratory support postoperatively; active or untreated asthma; allergic reaction to ibuprofen, acetaminophen, hydrocodone, and/or anesthesia, consumption of 3 or more alcoholic drinks each day and/or has a history of alcoholism; history of drug or alcohol abuse (excludes marijuana use); family history of drug or alcohol abuse in a rst degree relative; has had no more than one opioid prescription lled within the past 12 months; currently pregnant or lactating. In addition, individuals currently taking any of the following medications: CYP3A4 inhibitor, such as macrolide antibiotics (e.g., erythromycin), azole-antifungal agents (e.g., ketoconazole), and protease inhibitors (e.g., ritonavir), which may increase plasma concentrations of hydrocodone bitartrate and acetaminophen and prolong opioid adverse reactions, and which may cause potentially fatal respiratory depression or CNS depressants (including benzodiazepines), are excluded.
Participants are free to withdraw from participation in the study at any time upon request.
An investigator may discontinue an individual's participation in an intervention or withdraw an individual from the study if: Exparel (bupivacaine liposome injectable suspension) is used during the extraction procedure, the participant has a serious adverse event requiring hospitalization, any clinical adverse event (AE), laboratory abnormality, or other medical condition or situation occurs such that continued participation in the study would not be in the best interest of the participant; or the participant meets an exclusion criterion (either newly developed or not previously recognized) that precludes further study participation.
Who will perform the informed consent procedure? {26a} Consent is obtained by research coordinators at each clinical site. They are trained in human subjects protection, Good Clinical Practice and the OARS study protocol and procedures. The coordinator explains the research study to the potential participant undergoing mandibular 3 rd molar extraction and answer any questions that may arise. Extensive discussion of risks and possible bene ts of study participation is provided. A consent form describing in detail the study procedures and risks is given to the participant. Consent forms and recruitment materials are Institutional Review Board approved. Each participant signs the informed consent document prior to any study-related assessments or procedures. The participant may withdraw consent at any time throughout the course of the study. A copy of the executed consent form (either paper or electronic) is provided to the patient at the time of consent for their records.
Additional consent provisions for collection and use of participant data and biological specimens {26b} Not applicable.

Interventions
Explanation for the choice of comparators {6b} Two orally administered analgesics are being compared: (1) hydrocodone 5 mg/acetaminophen 300 mg and (2) ibuprofen 400 mg/acetaminophen 500 mg over a post-operative period.

Intervention description {11a}
Participants are directed to take 1 dose of either OPIOID or NON-OPIOID immediately after surgery and then 1 dose every 4-6 hours as needed for pain. If a dosage is not effective within an hour, a participant can take an additional dose, up to 6 total doses per day. In the event pain management remains insu cient, the subject is instructed to call the on-call surgeon. The on-call surgeon assesses the situation and may instruct the participant to take up to 2 additional doses or determine whether rescue medication (oxycodone 5 mg) or an emergency visit is required.

Criteria for discontinuing or modifying allocated interventions {11b}
Subjects who require rescue medication due to insu cient pain management with study analgesic are prescribed oxycodone 5mg in addition to the study analgesic. In addition, subjects who have adverse reactions which cannot be tolerated by the subject and/or subjects who have severe adverse reactions related to study analgesics are instructed to discontinue use of study analgesics.

Strategies to improve adherence to interventions {11c}
During the informed consent process, a thorough discussion with potential participants ensures that participants understand trial expectations. Subjects are provided written instructions after surgery. Short Message Service (SMS) texts are sent each morning and evening during the post-operative period Subjects receive SMS texts during the post-operative period between surgery and the postoperative visit in the morning and evening. In addition, research coordinators call participants to answer any questions and remind them of the study protocol.

Relevant concomitant care permitted or prohibited during the trial {11d}
Long-lasting local anesthetics such as Exparel and Marcaine (bupivacaine) are not permitted to be used.

Provisions for post-trial care {30}
Six months after completion of the trial, subjects who obtain additional opioids are offered, free-of-charge, an opioid addiction screening counselling session. Table 1 summarizes the outcome measures . The primary outcome measures for this study are patient-perceived pain (76) levels observed for seven days following 3 rd molar extraction and overall satisfaction (77) with the management acute post-surgical pain.

Outcomes {12}
Secondary outcome measures include: (1) frequency and magnitude of adverse events (77) , (2) ability to sleep (78) and pain interference (79) with normal daily activities, (3) potential diversion de ned as tablets remaining in households and returned at the follow-up visit, and (4) number of opioid prescriptions within 6 months after study analgesic use. A question from the PTSS was selected because it provides an overall rating of quality of sleep.
Pain Interference (Daily Function): The PROMIS Short 6b was selected because it is a standard NIH measure of pain interference and can be recorded during the post-operative visit.
Future Opioid Seeking Behavior: The PDMP is accessed; six months was selected because it is the maximum followup time which could be completed within the study time-frame.
Potential Diversion: Participants are instructed to bring the pill bottle and unused capsules to the follow-up appointment. Participants report whether they have experienced each adverse event and rate their sleep ability and pain interference in a daily electronic diary using a REDCap application developed for electronic phones and tablets. Expected Duration of Subject Participation: Subjects participate in the study for the time between their impacted 3 rd molar extraction procedure until their post-operative visit which normally occurs approximately 1 week later (9 days +/-5 days). Following state laws and regulations, the PDMP is retrospectively queried 6 months post-surgery to ascertain information about any opioid-containing prescriptions that may have been lled during this post-surgery time period. Figure 2 shows the participant timeline.
Sequence of Procedures and Duration of Study Period: Table 2 provides the schedule of events. To address the possible heterogeneity in pain management and side events in men and women due to sex differences and differences in pain tolerance, subgroup analysis in men and women will be performed separately to ensure the hypothesized differences between the analgesic groups remain.
Sample Size and Power: To test if the non-opioid analgesics are non-inferior to opioid analgesics for pain management (Hypothesis 1a), we chose a small and clinically non-signi cant difference d=1.0 as the non-inferiority margin and assumed the maximum standard deviation (SD) of the daily pain intensity to be 3.6, based on data from prior studies (42,(80)(81) . With 90% power and alpha=0.625% (one-sided, after Bonferroni correction for 4 tests of pain experience for day of surgery, day 1 and day 2 post surgery and last day of follow-up) (82) , we need at least 370 participants per group to test non-inferiority of non-opioid analgesics . To account for 15-20% loss of follow-up and missing data, and other factors not included in the sample size estimation, we will recruit 1800 participants with 450 participants x 2 analgesic groups x 2 gender subgroups (men and women). Power for subgroup analysis: With n=370/group (after attrition/missing data, etc.), non-inferiority margin d=1.0 and alpha=0.3125% (one-sided), we have 85% power to test non-inferiority of non-opioid analgesics in men and women separately.
To compare patient satisfaction (Hypothesis 1b), we estimated the minimal detectable difference in proportions of positive rating between NON-OPIOID vs. OPIOID for n=740/analgesic group (entire sample) and n=370/analgesic group (subgroup) analyses, after attrition/missing data. Assuming the proportion of positive rating (extremely satis ed and satis ed) for NON-OPIOID in our study is about 82%, similar to data in Daniels et al., (43) our study has 90% power to test a minimal difference of 7% (82% vs. 75%) for n=740/analgesic group (entire sample, 2-sided alpha=5%) and 11% (82% vs. 71%) for n=370/analgesic group (subgroup analysis, 2sided alpha=2.5%), in comparing proportion of positive rating of patient satisfaction. For continuous measures (e.g., mean normaldaily-function ratings and the sleep quality scores (Aim 2)), our study has 90% power to test a small effect size, Cohen's d of 0.20, for the entire sample (2-sided alpha=1.25%, after Bonferroni corrections for 4 tests for day of surgery, day 1 and day 2 post surgery and last day of follow-up) and Cohen's d=0.30 for the subgroup analysis (2-sided alpha=0.625%).
Consideration of within-site correlation: In the sample size and minimal effect size estimations, we tentatively assumed all individuals are independent. However, in our study, the randomization will be performed within each site. When the correlation between responses among participants within the same site is positive, randomizing participants within site can increase power: Consider a site of 2m participants and each patient is randomized to one of two treatments at a 1:1 ratio. Assume that the within-site correlation is r>0 and denote the standard deviation of each individual's response by σ. The variance of the mean difference in the responses between two treatments is 2σ 2 /m*(1-r). It is smaller than 2σ 2 /m, the variance of the same difference when individuals are mutually independent.
The smaller variance implies greater e ciency/power. Therefore, we expect our estimated sample size and minimal effect size to be more conservative than they should have been.

Recruitment {15}
Recruitment Strategies: Each clinical site recruits study participants as they report to the oral and maxillofacial surgery or general dentistry clinic for 3 rd molar extraction consultation or surgery appointments. Each site has study staff who review the clinic schedule at least each week for potential study candidates. When patients report to the clinic for a 3 rd molar extraction consult or procedure, study staff ascertain interest in study participation if at least one partial or full bony mandibular 3 rd molar extraction is planned. Sites advertise for the study by word of mouth and/or advertising through yers, social media, e-mail and/or ads in local/college newspapers.

Assignment of interventions: allocation
Sequence generation {16a} Participants were randomized to either the opioid or non-opioid analgesic group at the 1:1 ratio, strati ed by gender at each site. Sitespeci c pre-determined random number sequences, with a block of 4 containing 2 opioid and 2 non-opioid assignments in a random order, were generated by staff at Data Management Statistical Analysis Core (DMSA) using R software. The randomization code was generated, and labels are created during the preparation phase when the treatment packets are prepared so that each packet is prepared and labeled with the packet identi cation number according to the randomization sequence. Complete randomization code for each site is stored in REDCap and only the DMSA and Clinical Protocol Coordinating Core (CPCC) staff have access to it.

Concealment mechanism {16b}
Existing FDA approved caplets are placed into opaque capsules and then lled with powder ll. Study analgesics are provided in 2 different size/color capsules (Table 4). Capsule 1 is a brown AA capsule and contains either the hydrocodone/acetaminophen for OPIOID participants or ibuprofen for OPIOID participants. Capsule 2 is a white 00 capsule containing either a placebo for OPIOID or acetaminophen for NON-OPIOID. Capsule 1 for OPIOID and NON-OPIOID is manufactured with a similar weight to the extent possible.
Similarly, capsule 2 of OPIOID and NON-OPIOID is manufactured with a similar weight to the extent possible. The study Instructions for the taking the analgesic are the same for both the opioid and non-opioid cohorts, ensuring the patient, surgeon and site nurse coordinator are blinded. Unblinding is only to occur in the case of a medical emergency. If a medical emergency arises, clinical site personnel contact the Clinical Protocol Coordinating Chief who can perform the unblinding by looking up the subjects group assignment in the REDCap group assignment project, which is only available to the DMSA and the CPCC staff.

Data collection and management
Plans for assessment and collection of outcomes {18a} All data for this study is being electronically collected via REDCap forms and surveys, electronic medication bottles and activity monitors.
REDCap forms are being administered by study personnel using an iPad pro tablet. Data collected include eligibility criteria; demographics; pre-operative patient-reported data (i.e. pre-operative pain, pre-operative sleep ability, and pre-operative pain interference), surgical procedure data (i.e. Surgical Treatment Case Report, including: teeth extraction, type of impaction, extraction di culty, length of surgery, pharmaceuticals used) and post-operative patient-reported data (i.e. post-operative pain, post-operative sleep ability, post-operative pain interference, complications, overall satisfaction). All data collection instruments have been approved by the Rutgers University Institutional Review Board (IRB) and can be found in the electronic essential document binder for the OARS study. These forms are available upon request to the corresponding author.
REDCap surveys are used to capture subject eDiary entries. For these entries, a text message or email is sent to a participant's cell phone in the morning and evening during the post-operative period with the link to the REDCap surveys. Participants click on the link to access the surveys and make their eDiary entries directly through their cell phones.
SMRxT electronic medication bottles by SMRxT Inc. record the date and time each capsule is removed from the bottle. This provides the number of capsules removed each day and ultimately, the total number of capsules removed during the study period, enabling the determination of number of capsules available for diversion at the end of the post-operative period.
ActiGraph wGT3X-BT by ActiGraph, LLC sleep/activity monitors record active calories, total sleep minutes and deep sleep minutes.

Plans to promote participant retention and complete follow-up {18b}
To minimize loss of study participants and/or incomplete data collection, the following activities take place: An electronic message is sent to the participant each morning and each evening with a link to the e-diary and reminder to complete their e-diary entry.
If the eDiary entry is not made within 2 hours, an automatic text message reminder is sent up to two times Research coordinators contact participants by phone to reinforce adherence to study protocol. Rutgers University's Identity Management system, responsible for the enforcing of password complexity requirements and provides a centralized mechanism for the revocation of user access to integrated systems.
Users rights within REDCap are only assigned to REDCap projects, instruments, and records that are required for their role within the study. Further restrictions are put in place to ensure that study staff at each of the ve sites are granted access to data collected only within that site.
REDCap provides built-in audit trail functionality which logs all user activity including, but not limited to, page views, exporting data, entering data, and viewing or modifying elds.
SMRxT: Data collected via the SMRxT electronic pill bottles is stored on the SMRxT Nomi platform, transferred via cellular technology from the SMRxT bottle to Nomi. Nomi is directly integrated with REDCap enabling the download of SMRxT data.
ActiGraph: As part of the subject kit return to the CPCC process, the ActiGraph data is downloaded. Any issues with the download are identi ed. If the le does not successfully download, the ActiGraph is taken out of circulation and returned to the company for trouble shooting and data recovery.
Barcoding: Barcodes are used extensively throughout the study to label patient materials. Barcodes are generated at the Rutgers Core site using BarTender by Seagull Scienti c. Patient materials are then scanned into REDCap using a barcode scanner attached to dedicated study workstations to reduce human error when entering unique identi ers such as serial numbers or patient IDs into those systems.
Prior to enrolling the rst participant in the study, a disaster drill requiring system recover from a backup was successfully completed.
Data Validation: Whenever possible, structured responses and validation rules have been programmed requiring study participants and research staff to enter valid responses only (i.e., check a yes or no box, select from a list of possible answers such as a Likert scale, date has the format of a date -mm/dd/yy and falls within a certain range,). Branching logic in REDCap is used when respondents must provide answer to follow-up questions.
Stopping Logic: The REDCap system has also been designed to not allow research personnel to proceed with enrolling participants unless the consent process has been completed properly, inclusion criteria has been met and no exclusion criteria are present by blocking the research coordinator from entering the participant's mobile phone number or e-mail address. This stopping logic is documented within the REDCap forms.
Calculation Aids: In order to assist the research coordinators in determining whether visits are occurring in the allowable windows, the electronic data capture (EDC) system has been programmed to calculate the number of days between Visit 0 and Visit 1 and between Visit 1 and Visit 2. The EDC system has also been programmed to calculate the age of the participant, allowing the research coordinator to determine if the participant is at least 18 years old.

Con dentiality {27}
All identifying information (name, address, telephone numbers, date of birth) is stored in REDCap. Only authorized study staff who have the need to see this information have privileges to access the REDCap project. When staff leave the study, access privileges are removed. Each quarter, a review is done to verify that only active study staff have access privileges.
Deletion of Identifying Information: As per Rutgers University policy, once the trial is completed, all identifying information will only be stored for six years. After, identifying information will be deleted.
Database for Public Release: Only de-identi ed information will be released to the public upon request to the corresponding author.
Plans for collection, laboratory evaluation and storage of biological specimens for genetic or molecular analysis in this trial/future use

{33}
Not applicable Statistical methods

Statistical methods for primary and secondary outcomes {20a}
All statistical analyses will be performed on an intent-to-treat basis. To account for the repeated measures design and the correlations within study site, generalized linear mixed model (GLMM) analysis , (83) including mixed model and random effects logistic regression analyses, will be used to analyze the primary and secondary outcomes. In these analyses, participants and sites will be treated as random effects, where appropriate. For each test of an outcome, we de ne the statistical signi cance by p<0.05. Bonferroni correction will be applied for multiple testing, where appropriate.

Aim1 -Pain experience and Patient satisfaction
To test Hypothesis 1a, mixed model analysis will be used to model the pain intensity scores as a function of analgesic groups (nonopioid vs. opioid), day (day of treatment, day 1 and day 2 post surgery, and last day of follow-up) and day x analgesic group interactions as xed effects. Participants and sites will be adjusted using nested random effects to account for the possibly stronger within-patient than within-site correlations in the data. Variables such as, but not limited to, age, gender, and treatment variation (e.g., type of anesthesia, administration of a post-surgical steroid, di culty of surgery, time to complete surgery, etc.) will be controlled as covariates in the statistical model. Numbers of analgesic pills and duration between the rst and last doses taken during the day will also be controlled as time-dependent covariates in the statistical model as sensitivity analysis. Mean differences in pain intensity scores between NON-OPIOID and OPIOID groups (μ NonOpioid,day -μ Opioid,day ) on the day of treatment, averages from the day of treatment to each of the next two days, and the last day of follow up will be compared using linear contrasts. To test the non-inferiority of the non-opioid analgesics H 0 : (μ NonOpioid,day -μ Opioid,day )≥d vs. H 1 : (μ NonOpioid,day -μ Opioid,day )<d, we chose a small and clinically nonsigni cant difference d=1.0 as the non-inferiority margin .
(80-81,84-85) The non-inferiority will be assessed using the one-sided 99.375% (86) or two-sided 98.75%, (after Bonferroni correction for 4 tests of pain experience) con dence interval (CI) of μ NonOpioid,dayμ Opioid,day . If this CI completely lies below d, we then conclude the non-inferiority of the non-opioid analgesics; if it completely lies below 0, we then conclude the (statistical) superiority of the non-opioid analgesics at the 1.25% level (two-sided). (86-87). Summary statistics of numbers of pills per day that participants have taken during the entire period from the day of treatment until post-op will also be reported.
As a secondary outcome, percent of participants removed from the study due to the need for a rescue medication will be compared between NON-OPIOID and OPIOID groups using the random effects logistic regression analysis, with study site as the random effect.
To test Hypothesis 1b, we will calculate the proportion of each satisfaction category (e.g., extremely satis ed, satis ed, dissatis ed and extremely dissatis ed) and apply the chi-square test to compare between NONOPIOID vs. OPIOID. Generalized logistic regression analysis will be used to compare the distribution of satisfaction between NON-OPIOID vs. OPIOID, controlling for covariates. Patient satisfaction is a one-time measurement (asked on the day of post-op visit). Only study site will be adjusted as a random effect in the generalized logistic regression analysis.
Aim 2 -Adverse-Effects, Daily Function and Sleep Quality, and Opioid Seeking Behavior Adverse Events. The rate of the occurrence of adverse events will be compared using the random effects Poisson model with the total number of adverse event occurrences from each patient as the dependent variable, and the analgesic group (NON-OPIOID vs. OPIOID) as the independent variable ( xed effects). If patient follow-up time varies, we will include the person days as an offset term.
Covariates such as, but not limited to, age, gender and treatment variation, etc., will be controlled in the statistical models as xed effects. The same analyses will be repeated to compare the rate for each adverse event.
Percent of participants having to leave the study due to adverse events will also be compared using random effects logistic regression analysis as a secondary outcome. Site will be adjusted as a random effect in the statistical models.
Daily Function and Sleep Quality. Mean normal-daily-function and sleep quality ratings will be compared between NON-OPIOID vs. OPIOID using mixed model analysis.
Future Opioid Seeking Behavior. We hypothesize that participants receiving opioid prescriptions to help manage acute pain are more likely to receive at least one additional opioid prescription within 6 months. This outcome will be measured through a PDMP check at the 6 month point. Summary statistics and random effects logistic regression analysis will be used to compare the percent of participants lling opioid prescriptions (yes/no) within 6 months post-surgery between OPIOID and NON-OPIOID groups.
Drug diversion. We anticipate that participants receiving 5 days of opioid containing analgesics will have capsules remaining after their acute pain episode has been resolved. Descriptive statistics will be calculated and percent of participants who have capsules remaining and the number of capsules remaining will be reported.

Aim 3 -Clinical Protocol/Decision Support Tool.
To address Aim 3, we will develop an optimal (88-90) rule (model), as a function of patient characteristics, to recommend which analgesic to prescribe to maximize a patient's overall satisfaction. Speci cally, our sample will be divided into a training set and a testing set at 1:1 ratio. We will develop the optimal rule (89)(90) in the training set and validate it in the testing set, using the methods of Xu et al. (88) .

Interim analyses {21b}
Other than reviewing and summarizing adverse and serious adverse events, there are no plans for interim analyses.

Methods for additional analyses (e.g. subgroup analyses) {20b}
Heterogeneity and Subgroup Analyses: To address the possible heterogeneity in pain management and side events in men and women due to sex differences and differences in pain tolerance, we will perform subgroup analysis in men and women separately to ensure the hypothesized differences between the analgesic groups still remain. All the statistical analyses described above will be repeated for men and women respectively as subgroup analyses.
Sensitivity analyses -Site effect: To assess if the differences in NON-OPIOID and OPIOID varies by study site, we will perform sensitivity analysis by either adding interaction terms with site (i.e., site x analgesic groups, site x day and site x day x analgesic groups) to the statistical models or using strati ed analysis (strati ed by site).
Methods in analysis to handle protocol non-adherence and any statistical methods to handle missing data {20c} To assess the impact of missing data when we compare the outcomes between OPIOID and NON-OPIOID, sensitivity analyses assuming not missing-at-random (nMAR) and/or a mix of nMAR and missing-at-random (MAR) will be performed using multiple imputation procedures (91)(92)(93) to provide a spectrum of possible treatment effect between OPIOID and NON-OPIOID, helpful for describing plausible treatment effects, in the presence of missing data. We will also explore methods to model the missingness mechanism and apply the methods of selection models (94) or use the pattern-mixture models such as the control-based pattern imputation approach, or the tipping-point approach to handle missing data. (95)(96)(97)(98) Plans to give access to the full protocol, participant level-data and statistical code {31c} The full protocol and de-identi ed participant level data will be available to the public via a written request to the principal investigator.
Oversight and monitoring Composition of the coordinating centre and trial steering committee {5d} Figure 3 contains the organizational chart for the OARS study. There are two coordinating centers. The rst, appearing in blue, is the CPCC. This core is responsible for maintaining the protocol, manual of procedures, investigational product (IP) and all materials associated with the trial. In addition, the core is responsible for receiving subject kit orders from the clinical sites, processing the orders which includes assembling the subject kits, receiving all used subject kits from clinical sites, and destroying all materials including left-over investigational product.
The DMSA, which appears in red, is responsible for REDCap project development and upgrades, developing and regular reporting of trial quality metrics including enrolment and retention, randomization of subjects into the OPIOID or NON-OPIOID group, preparing quality management reports for the study team, the Data Safety Monitoring Board (DSMB), the FDA and the sponsor, locking the database at the end of the trial and completion of all statistical analyses.
There are two committees overseeing trial implementation. The rst, the Executive Committee is made up of the principal investigator (PI) and the 3 chiefs -the CPCC chief, the DMSC chief and the chief pharmacologist. The Executive Committee is responsible for overall project management and meets in conjunction with the Steering Committee. Should an emergency meeting be required due to a halting rule being reached or a serious adverse even occurring likely due to IP, the Steering Committee would review the vents to determine the appropriate action. In addition, the Steering Committee is responsible for approving all plans for manuscript submissions.
The Steering Committee consists of all Executive Committee members along with the site directors. The Steering Committee is responsible for the implementation of the protocol at each of the clinical sites. The Steering Committee meets monthly to review the study's quality metrics, review any adverse events, serious adverse events, protocol deviations and unanticipated problems. In addition, any changes to the protocol, manual of procedures, clinical quality management plan or data quality plan are discussed. IT staff, and research coordinators routinely attend the Steering Committee Meetings.
Composition of the data monitoring committee, its role and reporting structure {21a} There are two data monitoring committees which provide trial oversight in addition to oversight reporting to the FDA.
NIH-NIDCR has appointed a clinical monitor who is responsible conducting regular audits. The clinical monitor performs both blinded and unblinded monitoring visits and is responsible for ensuring protection of human subjects and the integrity of the research data, along with adherence to the study protocol. Blinded monitoring visits focus on the informed consent process and consent records, complete study participant records, event reporting, training of blinded study staff and maintenance of the projects essential documents including IRB records, the protocol, and the manual of procedures. Unblinded monitoring visits focus on investigational product, maintenance of blinding, training of unblinded study staff, and delegation of duties for unblinded staff. Reports are generated from each monitoring visit and shared with the sponsor and the DSMB.
The DSMB is responsible ensuring study participant safety. The frequency of reviews is dependent upon risk to the subjects. The DSMB consists of 5 members who are biostatistical, clinical decision support, treatment of addictive disorders, clinical dental pharmacology, and oral and maxillofacial surgery experts. The DSMB reviews recruitment and retention, adverse events. The DSMB periodically reviews accumulated study data for participant safety, study conduct and progress and, if required, make recommendations concerning the continuation, modi cation, or termination of the trial. In addition to the NIH-NIDCR appointed monitor and DSMB, the OARS study team performs weekly reviews of all subject records completed that week and performs a quarterly quality management review. This review enables de ciencies to be identi ed promptly with queries sent to the appropriate site personnel to address identi ed de ciencies.
A quarterly management review is conducted by the PI, CPCC Chief and DSMB Chief. During these reviews 10% of completed records are randomly selected and reviewed for de ciencies. All adverse events, serious adverse events, protocol deviations and unanticipated problems are reviewed along with the outcome of recommendations developed from the previous review. Adherence to the protocol and manual of procedures is checked along with the maintenance of all study essential documents.
Plans for communicating important protocol amendments to relevant parties (e.g. trial participants, ethical committees) {25} All protocol amendments are approved by the Rutgers IRB prior to implementation. Changes are communicated to study staff and the sponsor via a numbered memorandum and reviewed at monthly study staff meetings. The current protocol version along with archived versions of the protocol are maintained in the essential documents electronic binder which is available for review by study staff, monitors and sponsor.

Dissemination plans {31a}
This study complies with all applicable NIH Data Sharing Policies. (100) Dissemination at Scienti c Meetings: Presentations at scienti c meetings will be delivered to assist in dissemination of results as soon as possible when nal results pertaining to the primary variables are available. Meetings at which presentations will be made include, but will not be limited to, American and International Association of Dental Research (AADR and IADR), the American Dental Association (ADA), and International Association for the Study of Pain (IASP) and addiction meetings. NIH grant support will be acknowledged during all presentations.
Publication and Authorship Policies: Findings will be published in peer-reviewed journals. Journals selected will be indexed in PubMed. ICMJE guidelines will be adopted and followed in determining authorship. Accepted manuscripts will be submitted to PubMed Central as per NIH policy. Final versions of the peer reviewed manuscripts will be made available to the public, generally within 3 months but no later than 12 months after the o cial date of publication. NIH grant support will be acknowledged in all publications.
NIH Public Access Policy: The NIH Public Access Policy requires scientists to submit nal peer-reviewed journal manuscripts that arise from NIH funds to PubMed Central immediately upon acceptance for publication. This ensures that the public has access to the published results of NIH funded research.

Discussion
Dentists often prescribe opioids to their patients to manage acute post-surgical pain, with hydrocodone/acetaminophen being the most commonly prescribed opioid combination. With millions of opioid prescriptions written each year and increasing evidence that adolescents taking opioids are at increased risk of addiction as adults, it is vitally important to develop the best evidence for managing acute post-surgical pain.
Of note is the pragmatic nature of this trial (Table 5). First, unlike most pain studies which follow patients for only a limited number of hours post-surgery, our participants are followed for the normal post-operative period of one to two weeks covering the entire acute pain episode. Second, randomization is performed within each site, strati ed by gender to ensure that results apply equally to men and women with subgroup analysis to be performed. Third, other than the rst dose which is required enabling patients to "get ahead of the pain", we propose using medication the way it is prescribed in clinical practice today. Lastly, our primary study outcome is pain experience and satisfaction with managing pain rather than complete pain relief (total elimination of pain after the pain has risen to a critical level.) Dental providers are uniquely positioned to take a leadership role in mitigating the national opioid crisis by reducing the number of opioids prescribed to manage acute pain following common dental procedures. (99) This clinical trial will provide key information to impact their prescribing habits. JFY is the chief of the clinical protocol coordinating core. She assisted in protocol development and is principal author of the Manual of Procedures. She assisted in the writing of this manuscript.
SEL is the chief of the statistical and data management core. She performed sample size analysis and designed the statistical analysis plan. She was involved with manuscript review and re nement.
PJD is the chief of pharmacology. He contributed to study design, assisted in protocol development, reviewed the Manual of Procedures and was participated in manuscript review and re nement.
MM is a clinical site director. He assisted in contributed to study design, protocol re nement and participated in manuscript review and re nement.
HM is a clinical site director. He contributed to study design, assisted in protocol re nement and participated in manuscript review and re nement.
GW is a clinical site director. He contributed to study design, assisted in protocol re nement and participated in manuscript review and re nement.
BW is a clinical site director. He contributed to study design, assisted in protocol re nement and participated in manuscript review and re nement.
VZ is a clinical site director. He contributed to study design, assisted in protocol re nement and participated in manuscript review and re nement.
DF Is a scienti c advisor. He contributed to study design, assisted in protocol re nement and participated in manuscript review and re nement.
All authors read and approved the nal manuscript.

Funding {4}
Research reported in this manuscript is supported by National Institute of Dental and Craniofacial of the National Institutes of Health awards under award numbers UG3DE028860 and UH3DE028860. The content is solely the responsibility of the authors and does not necessarily represent the o cial views of the National Institutes of Health.

Availability of data and materials {29}
This study complies with all applicable NIH Data Sharing Policies. (100) NIH Policy on the Dissemination of NIH-Funded Clinical Trial Information: The study is a clinical trial and complies with the NIH policy that establishes the expectation that all investigators conducting clinical trials funded in whole or in part by the NIH ensure that these trials are registered at ClinicalTrials.gov, and that results of these trials are submitted to ClinicalTrials.gov.
This clinical trial, was registered on the ClincialTrials.gov by Cecile A. Feldman, DMD, the study principal investigator, no later than 21 days after enrollment of the rst subject. Data to be posted at time of initial registration include descriptive information, recruitment information, location and contact information and administrative information. Final results will be posted no later than 12 months after the trial's primary completion date and will include participant ow, demographic and baseline characteristics, outcomes and statistical analyses, adverse events, the protocol, statistical analysis plan and administrative information.
The PI will share the individual non-identi ed subject data no later than acceptance for the publication's main ndings from the nal data analysis or 18 months after completion of the study, whichever is earlier. As NIDCR does not have a data repository at the initiation of this study, data will be posted on the Rutgers designated clinical trials data warehouse. (If NIDCR has a clinical trial data repository at the time of study, completion, data will be posted on the NIDCR data repository.) Data to be posted will include the study protocol, reference to study publication of primary outcome variable, the data set in both SAS and ASCII formats, data dictionary and study speci c de-identi cation notes.
Upon written request to the PI, survey instruments or other materials developed for use during the clinical trial will be made electronically available to other researchers.
Public use datasets will be available at least 18 months after study completion. The Data Management and Statistical Core Chief will be responsible for developing the detailed public dataset documentation by the end of UH3 Year 2. Included will be a description of the data format, the data elements and the data dictionary. Public use datasets will be available upon request.  OARS Organization Chart