predicting and disrupting the opioid addiction cycle€¦ · source: ny times article "short...
TRANSCRIPT
The opinions expressed in this presentation are those of the speaker. The International Society and International Foundation disclaim responsibility for views expressed and statements made by the program speakers.
Predicting and Disrupting the Opioid Addiction Cycle
Ted BorgstadtChief Executive Officer
TrestleTree LLOFayetteville, Arkansas
Bradley Martin, Pharm.D., Ph.D.UAMS College of Pharmacy
Little Rock, Arkansas
Julia NicholsonPrincipal
JN ConsultingFolsom, California
4C-1
AGENDA
• HOW BAD IS THE PROBLEM?
• VALUABLE INSIGHTS FROM 20 YEARS OF RESEARCH
• PROACTIVE STRATEGIES TO ADDRESS THE PROBLEM
• Q & A
4C-2
HOW BAD IS IT?
• LEADING CAUSE OF DEATH FOR PEOPLE UNDER 50
• 2% OF DEATH = 1 IN 50SOURCE: NY Times article "Short Answers to Hard Questions About the Opioid Crisis" by Josh Katz,
August 2017 citing the CDC and National Center for Health Statistics
• 97 MILLION PEOPLE TOOK PRESCRIPTION PAINKILLERS IN 2015 BUT 12 MILLION WITHOUT A DOCTOR PRESCRIBING
• OVER 2 MILLION PEOPLE ESTIMATED TO HAVE A PROBLEM WITH OPIOIDS
4C-3
WIIFM
• “ECONOMIC BURDEN”
• INDIRECT EFFECTS
• PROACTIVE STEPS
4C-4
OPIOID USE TRAJECTORIES
WHAT WE HAVE LEARNED FROM BIG DATA
4C-5
TERMINOLOGY:Abuse: Self-administration of medications to alter one’s state of consciousness (“get high”)
Addiction: A primary, chronic, neurobiological disease, with genetic, psychologic, and environmental factors influencing its development and manifestations. Addiction is characterized by 4 C’s – behaviors that include one or more of the following: impaired control over drug use, compulsive use, continued use despite harm, and craving
Diversion: Redirection of a prescription drug from its lawful purpose to illicit use; can be done with criminal intent
Misuse (noncompliant use): The intentional or unintentional use of a prescribed medication in a manner that is contrary to directions
4C-6
TERMINOLOGY:Physical dependence: A state of adaptation manifested by a drug class-specific withdrawal syndrome that occurs by abrupt cessation of a drug, symptoms generally the opposite of the desire drug effect, and by adaptation so that increasing doses of a drug is needed to achieve the same desired effect.
Tolerance: A state of adaptation in which exposure to a given dose of a drug induces changes that result in diminution of one or more of the drug’s effects over time
Withdrawal: A variety of unpleasant symptoms (e.g., difficulty concentrating, irritability, anxiety, anger, depressed mood, sleep disturbance, and craving) that occur after use of an addictive drugs is reduced or stopped
4C-7
CHRONIC PAIN AND PRESCRIPTION OPIOIDS
• 11% of Americans experience daily (chronic) pain
• Opioids frequently prescribed for chronic pain
• Primary care providers commonly treat chronic, non-cancer pain
o account for ~50% of opioid pain medications dispensed
o report concern about opioids and insufficient training
4C-8
• Adopted guidance to more
aggressively monitor and treat
paint 1999-2000
Veterans Health Administration
• 1995 Pain as fifth a vital sign
• 90% of American Pain
Society Funding from
PhRMA
American Pain Society
MANAGING PAIN
4C-9
OPIOID RELATED ARTICLES AND LETTERS
4C-10
N Engl J Med 2017; 376:2194-2195
4C-11
EARLY ASSOCIATION BETWEEN CUMULATIVE OPIOID DOSE AND OPIOID DEATH
4C-12
4C-13
4C-14
Use Big ImagesTo Show Ideas
OPIOID USE COMPARED TO PALLIATIVE CARE NEEDS
4C-15
RAPID INCREASE IN DRUG OVERDOSE DEATH RATES BY COUNTY
4C-16
SHARP INCREASE IN OPIOID PRESCRIPTIONS INCREASE IN DEATHS
4C-17
COMPARISON OF POTENTIALLY LETHAL DOSES
ILLICIT OPIOIDS
PHOTO: Kensington Police – PEI Canada - 2017
4C-18
3
2
1
0
4
5
6
7
2004 2006 2008 2016
Dea
ths
per1
00,0
00po
pula
tion
RISE IN OPIOID DEATHS
Over 350,000 people have died from an opioid overdose since 1999
3 Waves
SOURCE: National Vital Statistics System Mortality File
ILLICIT OPIOIDSHeroin
2000 2002
PRESCRIPTION OPIOIDSNatural and semi-synthetic
opioidsMethadone
2010 2012 2014
OTHER – LIKELY ILLICIT FENTANYLSynthetic Opioids
Excluding Methadone
Overlapping, Entangles but Distinct Epidemics
4C-19
STAT forecast: Opioids could kill nearly 500,000 Americans in the next decadeBY MAX BLAU @MAXBLAUJUNE 27, 2017
10 YEAR PROJECTIONS OF OPIOID RELATED DEATHS
4C-20
LOOKING BEHIND OPIOID OVERDOSE DEATHS
ForeveryONEperson who diedthere were
15 Disorder involving heroin people reported having a substance use
In 2016,more than42,000Americans died fromoverdoses involvingprescription or illicitopioids.
people reported having a substance use disorder Involving opioids
people reported misusingprescription opioids in the past year
people reported usingprescription opioids in the past year
41
273
2,174
4C-21
DRIVERS OF CHANGES IN LIFE EXPECTANCY 2000-201512 Leading causes of Death and Drug Poisoning Deaths
Copyright 2017 American Medical Association. All Rights Reserved.
4C-22
4C-23
4C-24
4C-25
LONG-TERM CHRONIC OPIOID THERAPY DISCONTINUATION RATES FROM THE TROUP STUDY
• The study was designed to report chronic opioid therapy discontinuation rates after five years and identify factors associated with discontinuation.
• Commercially insured population (HealthCore plans)
• Publicly Insured population (Arkansas Medicaid)
4C-26
DAYS UNTIL OPIOID DISCONTINUATION BY MISUSE SCORE FOR ARKANSAS MEDICAID AND HEALTHCORE, 2001-2005
N = 23,419 and 6,848
4C-27
0
5
10
15
20
25
30
35
40
0 10 20 30 40 50Pro
babi
lity
of c
ontin
uing
use
in %
Number of days of first episode of opioid useOne year probability Three year probability
CONTINUED USE BY INTIAL DAYS OF THERAPY (N=1,294,247)
4C-28
CONTINUED USE BY NUMBER OF PRESCRIPTIONS
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8 10 12 14 16
Pro
babi
lity
of c
ontin
uing
use
in
%
Number of prescriptions in first episode of opioid useOne year probability Three year probability
4C-29
CONTINUED USE BY DAYS SUPPLIED OF FIRST PRESCRIPTION
0
5
10
15
20
25
30
35
40
45
50
0 5 10 15 20 25 30 35 40 45Pro
babi
lity
of c
ontin
uing
use
in
%
Days' supply of the first opioid prescription
One year probability Three year probability
4C-30
CONTINUED USE BY CUMULATIVE DOSE (MEQ) OF FIRST EPISODE
0
10
20
30
40
50
60
0 500 1000 1500 2000 2500
Pro
babi
lity
of c
ontin
uing
use
in
%
Cumulative dose in first episode of opioid useOne year probability Three year probability
4C-31
CONTINUED USE BY INTIAL OPIOID TYPE
0 5 10 15 20 25 30
Long Acting Opioids (6,588; 0.51%)
Tramadol (120,781; 9.33%)
Hydrocodone Short Acting (742,112; 57.33%)
Oxcodone Short Acting (219,224; 16.94%)
Schedule II Short Acting (14,877; 1.15%)
Schedule III-IV and Nalbuphine (190,665;14.74%)
Initial Opioid Type
Three year probability of continued use One year probability of continued use
4C-32
COMPARISON WITH ANIMAL MODELS
Schulteis, et.al. Pharmacology Biochemistry and Behavior. 2004
4C-33
• Opioids for 13 weeks
have high probabilities of
long term use (61.11%)
and nearly constant after
that
Tramadol appears to be used sometimes when clinicians are thinking of a long term analgesic strategy
• 1 or more days 6.00%
• 8 or more days 13.52%
• 31 or more days 29.85%
Each additional day of opioid therapy increases the risk of chronic opioid use starting with 3rd day
No clear thresholds based on weeks of initial opioid use or cumulative dose
THOUGHTS
4C-34
Explore the association between characteristics of the first opioid prescription, patient level factors, and the pain etiology on the
probability of opioid discontinuation among opioid naïve patients
4C-35
TIME TO OPIOID DISCONTINUATION BY PAIN ETIOLOGY
4C-36
• Medicaid, Long Acting Opioid, Chronic Pain, 3-4 days suppliedModerate Associations (0.6 < HR < 0.8)
• Women, Elderly, Living in South or West, All Mental Health Disorders Studied, Pain types except chronic pain, childbirth, and surgery, Benzodiazepine and Muscle Relaxant Use, High Daily Dose, Use of Oxycodone, Hydrocodone, Tramadol, Short Acting Schedule II
Modest Associations (0.8 < HR < 1.0)
LONGER OPIOID USE IS ASSOCIATED WITH:
Strong Associations (HR < 0.6)• Days supplied greater than 4
4C-37
HAZARD OF OPIOID DISCONTINUATION BY DAYS SUPPLIED INTERACTED WITH PAIN ETIOLOGY
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
3‐4 DAYS 5‐7 DAYS 8‐10 DAYS 11‐14 DAYS 15‐21 DAYS 22 OR MORE DAYS
Hazard Ratios
Days' Supply of First PrescriptionBURN TRAUMA AND SURGERY TRAUMASURGERY CHILDBIRTH DENTAL PROCEDURECHRONIC PAIN DIAGNOSIS NON‐CHRONIC PAIN OTHER INPATIENT ADMISSIONOTHER EMERGENCY ROOM VISIT
4C-38
Persons with mental health disorders and those prescribed muscle relaxants or benzodiazepines used opioids longer
• Persons prescribed 5-7 days are TWICE as Likely to Continue Opioids than persons prescribed 1-2 days
• Persons prescribed 11-14 days are THREE times as Likely to Continue Opioids
Days Supply of Initial Opioid is the STRONGEST Predictor of Long Term Opioid Use
The Effect of Days Supplied persisted across all pain etiologies
THOUGHTS
4C-39
• States are implementing initial days supply limits – usually 7 days
or less
Policy
• The Initial Days Supplied of Opioids is the strongest
modifiable factor prognostic for long term use
• Extra Caution for persons with mental health
disorders and those prescribed
muscle relaxants or benzodiazepines
ClinicalIMPLICATIONS
• Supports CDC guidelines limiting initial opioids to
3 days or less and rarely more than 7 days
• PBMs are implementing similar restrictions
4C-40
WHO’S AT RISK FOR OVERDOSE?• DEMOGRAPHICS
• SOCIOECONOMICS & GEOGRAPHY
• RISK FACTORS
oMeno35-54 year oldsoWhitesoAmerican Indians / Alaska Natives
oMedicaidoRural
oPatients receiving opioids from multiple prescribers and/or pharmacies
oPatients taking high daily doses of opioidsoDepression, substance use disorder, and prior
overdose
4C-41
PREDICTING OPIOID ADVERSE EVENTS
Rates are per 1,000 person years
UAMS Translational Research Institute
4C-42
PREDICTING OPIOID ADVERSE EVENTS
UAMS Translational Research Institute
4C-43
PREDICTING OPIOID ADVERSE EVENTS
0.00%
0.10%
0.20%
0.30%
0.40%
0.50%
0.60%
0.70%
0.80%
0.90%
1.00%
0
0.2
0.4
0.6
0.8
1
1.2
1 2 3 4 5 6 7 8 9 10
Predicted Risk Observed Overdose Rate
UAMS Translational Research Institute
4C-44
PREDICTING OPIOID ADVERSE EVENTS
UAMS Translational Research Institute
4C-45
• Medical Marijuana Legalization Appears to Modestly Decrease Opioid Prescribing and Chronic Opioid Abuse
• Chronic and High Risk Opioid Prescribing appears Highly Concentrated
HOT OFF THE PRESS
• Only 24% of Reproductive Age Women who use Opioids Chronically obtain Prescribed Birth Control
• Qualitative Study of Pharmacists and Physicians Attitudes towards opioid decisions
4C-46
Acknowledgements / Disclosers
Anuh Shah, Corey Hayes,
Xiaocong Li, Gary Moore
Access to the Data was supported
by the UAMS TRI (UL1TR000039)
NIDA (R01 DA 022560) Sullivan
Disclosure: TrestleTree
4C-47
PREVENTION DISRUPTING THE OPIOID ADDICTION CYCLE
BEFORE IT BEGINS
4C-48
“The opioid crisis is just awful and it is getting worse.”
Sarah Bacon, PhDDivision of Unintentional Injury PreventionCenters for Disease Control & Prevention
2018 Opioid Epidemic ForumWashington DCJuly 16, 2018
4C-49
“It will take at least 10 years to solve the opioid crisis.”
Theodore A. Christopher, MD, FACEPThomas Jefferson University Hospital
2018 Opioid Epidemic ForumWashington DCJuly 16, 2018
4C-50
LOOKING BEHIND OPIOID OVERDOSE DEATHS
ForeveryONEperson who diedthere were
15 Disorder involving heroin people reported having a substance use
In 2016,more than42,000Americans died fromoverdoses involvingprescription or illicitopioids.
people reported having a substance use disorder Involving opioids
people reported misusingprescription opioids in the past year
people reported usingprescription opioids in the past year
41
273
2,174
4C-51
CDC RESPONSE
CDC VISION
Prevent Opioid-Related Harms & Overdose Deaths
North Star
4C-52
ConnectingCommunity
Infrastructure
COORDINATEDPrevention and
ResponseActivities
4C-53
• State Legislative Efforts
• Health Plans
• Hospital Systems
• Pharmacy Benefit Management
• Retail Pharmacies
CURRENT EFFORTS
4C-54
STATE LEGISLATURES
4C-55
MEMBERS• Identify a member once an
opioid prescription has been written
• Reach out to the member to educate them on the serious risk of taking an opioid
• Expand access to alternative pain management therapies
HEALTH PLANS
PHYSICIANS / PROVIDERS• Identify the number of opioid
prescriptions each physician has written in comparison to their peers via a scorecard
• Educate the physician on the alternative pain management options covered by the health plan
4C-56
MEMBERS• Identify a member once an
opioid prescription has been written
• Reach out to the member to educate them on the serious risks of taking an opioid
• Disposal of unused opioids
HOSPITAL SYSTEMS
PHYSICIANS / PROVIDERS• Identify the number of opioid
prescriptions each physician has written in comparison to their peers
• Educate the physician on pain management strategies and techniques
• Engage new policy & procedures for Emergency Department opioid utilization
4C-57
MEMBERS• Identify a member once an
opioid prescription has been written
• Reach out to the member to educate them on the serious risks of taking an opioid
• Monitor doctor shopping
• Repetitive emergency department prescription
PHARMACY BENEFIT MANAGEMENT COMPANIES
PHYSICIANS / PROVIDERS• Identify the number of opioid
prescriptions each physician has written in comparison to their peers
• Educate the physician on step therapy recommendations for pain management
4C-58
CUSTOMERS• Identify a member once an
opioid prescription has been received to be filled
• Counsel customer in the store to educate them on the serious risks of opioids
• Restrict the number of days supply of an opioid prescription to 7 days
• Disposal of unused opioids
RETAIL PHARMACY
PHYSICIANS / PROVIDERS
• Identify and contact the physicians that have written over a seven days supply and alert them to the restricted days supplied filled
4C-59
• Current Strategies from multiple health care sectors- Identify - Educate
ADDITIVE SOLUTIONS NEEDED
• Focus is on disrupting opioid abuse before the unintentional slippery slope impacts lives and increases avoidable cost
• Additive Disruptive Strategy for all health care sectors- Predict- Prevent
4C-60
• This crisis requires an additional approach focused on patient-level insights to predict opioid addiction risk and leverage behavior change methodology to preventopioid misuse and addiction at the individual level
PREDICTION OF OPIOID RISK
• Focus is on disrupting opioid abuse before the unintentional slippery slope impacts lives andincreases avoidable cost
• 323 risk variables (“data features”) spanning demographic pharmacy claims, and medical claims data, each individually weighted for an individual prior to being prescribed an opioid for the first time
4C-61
SLIPPERY SLOPE OF OPIOID ABUSE: 5+ DAYS
• 14% likelihood of continued opioid use 1 year after initial 7 day supply• 25% likelihood of continued opioid use 1 year after initial 11 day supply• No previous opioid use in the proceeding 6 months for ~1.3m patients studied
4C-62
• Breaking the cycle of abuse and addiction before it begins
• Helping people change tough health behaviors who do not want to change
• Behavior change intervention for high-risk individuals who are being prescribed an opioid for the first time, before a refill, pre-and post-surgery, post-ER visit, work comp injury, etc.
• Need a full contextual understanding of patient predisposition for opioid abuse, such as family, culture, finances, social, etc.
• Holistic approach to health behavior change allows influence on co-morbidities and foster referrals/access to all available resources
PREVENTION OF OPIOID RISK
4C-63
EXAMPLE: PRE- & POST-SURGERY
4C-64
EXAMPLE: WORK COMP PT
4C-65
FINAL THOUGHTS
• “The opioid crisis is just awful and it is getting worse.”
• Over 2 million people estimated to have a problem with opioids
• This crisis requires an additional approach focused on patient-level insights to predict opioid addiction risk and leverage behavior change methodology to prevent opioid misuse and addiction at the individual level
• Focus is on disrupting opioid abuse before the unintentional slippery slope impacts lives and increases avoidable cost
4C-66
QUESTIONS?
4C-67
Back Up Slides with Additional Methods and Data
4C-68
OPIOID WITHDRAWAL
4C-69
• Study Measures:
• Data: IMS Lifelink+ database for 2006-2015
• Subjects: Opioid naïve, cancer and substance
abuse free patients
DATA, SUBJECTS, KEY STUDY MEASURES
• Opioid Discontinuation at 1 and 3 years
• Opioid Prescription Characteristics
• Initial Days of Opioid Use
• Number of Opioid
prescriptions
• Cumulative Dose
• Type of opioid
• Initial Days Supply
of Opioid
4C-70
SUBJECTSTotal = 1,294,247 N / Mean (% / s.d.)Age 44.52 (14.56)Gender Females 698950 (54.00)Enrollment duration (years) 2.48 (2.04)
Pain Diagnoses Back Pain 226681 (17.51) Neck Pain 90352 (6.98) Head Pain 30123 (2.33) Joint Pain 389700 (30.11)
Patient Region South 476565 (36.74) Midwest 376520 (29.09) East 279595 (21.60) West 142698 (11.03) Missing/Other 19869 (1.54)Payer Type Commercial 866815 (66.97) Self‐Insured 387122 (29.91) Other / Unknown 40310 (3.11)
Opioid Episode Characteristics First Prescription >= 90 MEQ 89438 (6.91) First Prescription >= 120 MEQ 22895 (1.77) First Prescription of Long Acting Opioid 6588 (0.51) Duration of first episode 14.81 (65.00)
4C-71
Opioid naïve, cancer and substance abuse free patients
IMS Lifelink+ database for 2006-2015• 10% random sample (approximately 6.5M lives)
DATA AND SUBJECTS
• At least one OPR prescription between June 06 and Dec 14
• At least 6 months of continuous enrollment without an opioid prescription
prior to first opioid prescription
• At least 14 years of age
• Excluded:
• Any non-melanoma cancer, substance abuse diagnosis or
buprenorphine/naloxone prescription in the 6 month prior period
• Missing data on demographics (gender, region, age, payer)
4C-72
Initial Opioid Episode• Days supply of first prescription, average daily dose, opioid type
Opioid Discontinuation
STUDY MEASURES
Pain Etiology• (a) Trauma and surgery (b) Trauma (c) Surgery (d) Burn (e) Childbirth (f)
Dental (g) Chronic pain conditions (headache, back/neck pain, joint pain, neuropathic pain, fibromyalgia) (h) Other pain conditions (chest pain, abdominal pain, others) (i) Other inpatient admissions (j) Other Emergency Department visit (k) Unknown
Patient and System Characteristics• Age, Gender, Region, Year, Insurance Type (Medicaid, Medicare,
Commercial), Mental Health (mood, personality, adjustment, anxiety), Prior Benzodiazepine/Muscle Relaxant.
4C-73
Opioid naïve, cancer and substance abuse free patients
IMS Lifelink+ database for 2006-2015• 10% random sample (approximately 6.5M lives)
DATA AND SUBJECTS
• At least one OPR prescription between June 06 and Sep 15
• At least 6 months of continuous enrollment without an opioid prescription
prior to first opioid prescription
• At least 18 years of age
• Excluded:
• Any non-melanoma cancer, substance abuse diagnosis or
buprenorphine/naloxone prescription in the 6 month prior period
4C-74
Initial Opioid Episode• Continuous use of opioids with a gap no greater than 30 days• Duration (days and weeks)• Number of opioid prescriptions• Cumulative dose (expressed in morphine milliequivalents)
Opioid Discontinuation• At least 180 days without opioid use• Followed until loss of enrollment, study end date, or
discontinuation of opioids.
STUDY MEASURES
Initial Prescription• Days supply• Average daily dose (MME)• Prescription type
• Long Acting, Oxycodone Short Acting, Hydrocodone Short Acting, Other Schedule II Short Acting, Schedule III-IV, Tramadol
4C-75
Sensitivity analyses• Discontinuation required 90 instead of 180 opioid free days
• Initial opioid episode used 7 instead of 30 maximum gap
• Excluded patients whose initial prescription exceeded 90 MME
Kaplan Meier curves• Median time to discontinuation
• Probability of continued OPR use at one and three years
ANALYSIS
4C-76
SUBJECTSTotal = 1,294,247 N / Mean (% / s.d.)Age 44.52 (14.56)Gender Females 698950 (54.00)Enrollment duration (years) 2.48 (2.04)
Pain Diagnoses Back Pain 226681 (17.51) Neck Pain 90352 (6.98) Head Pain 30123 (2.33) Joint Pain 389700 (30.11)
Patient Region South 476565 (36.74) Midwest 376520 (29.09) East 279595 (21.60) West 142698 (11.03) Missing/Other 19869 (1.54)Payer Type Commercial 866815 (66.97) Self‐Insured 387122 (29.91) Other / Unknown 40310 (3.11)
Opioid Episode Characteristics First Prescription >= 90 MEQ 89438 (6.91) First Prescription >= 120 MEQ 22895 (1.77) First Prescription of Long Acting Opioid 6588 (0.51) Duration of first episode 14.81 (65.00)
4C-77
Unable to account for opioids that were paid for out of pocket or obtained illicitly
Unable to separate intentional and un-intentional long term chronic opioid use
Data do not capture pain intensity of duration
Did not account for the etiology of pain• Acute (post-op, trauma) or chronic pain conditions
LIMITATIONS
4C-78
Opioid naïve, cancer and substance abuse free patients
IMS Lifelink+ database for 2006-2015• 10% random sample (approximately 6.5M lives)
DATA AND SUBJECTS
• At least one OPR prescription between June 06 and Dec 14
• At least 6 months of continuous enrollment without an opioid prescription
prior to first opioid prescription
• At least 14 years of age
• Excluded:• Any non-melanoma cancer, substance abuse diagnosis or
buprenorphine/naloxone prescription in the 6 month prior period
• Missing data on demographics (gender, region, age, payer)
4C-79
Initial Opioid Episode• Days supply of first prescription, average daily dose, opioid type
Opioid Discontinuation
STUDY MEASURES
Pain Etiology• (a) Trauma and surgery (b) Trauma (c) Surgery (d) Burn (e) Childbirth (f)
Dental (g) Chronic pain conditions (headache, back/neck pain, joint pain, neuropathic pain, fibromyalgia) (h) Other pain conditions (chest pain, abdominal pain, others) (i) Other inpatient admissions (j) Other Emergency Department visit (k) Unknown
Patient and System Characteristics• Age, Gender, Region, Year, Insurance Type (Medicaid, Medicare,
Commercial), Mental Health (mood, personality, adjustment, anxiety), Prior Benzodiazepine/Muscle Relaxant.
4C-80
Cox Proportional Hazards Model• General Model
• Interacted Model
• Stratified Model
Kaplan Meier curves
ANALYSIS
• Pain etiology and days supplied
• Chronic vs Non-Chronic Pain
4C-81
4C-82
4C-83
TIME TO DISCONTINUATION
4C-84
4C-85
4C-86
PREDICTING OPIOID ADVERSE EVENTS Statistical Approach
UAMS Translational Research Institute
4C-87