MONITORING PROGRAMS ON OPIOID PRESCRIBING & USE REBECCA L. - - PowerPoint PPT Presentation
EFFECTS OF ROBUST STATE PRESCRIPTION DRUG MONITORING PROGRAMS ON OPIOID PRESCRIBING & USE REBECCA L. HAFFAJEE, JD, PHD, MPH ASSISTANT PROFESSOR OF HEALTH MANAGEMENT & POLICY UNIVERSITY OF MICHIGAN SCHOOL OF PUBLIC HEALTH ACADEMY HEALTH
REBECCA L. HAFFAJEE, JD, PHD, MPH ASSISTANT PROFESSOR OF HEALTH MANAGEMENT & POLICY UNIVERSITY OF MICHIGAN SCHOOL OF PUBLIC HEALTH ACADEMY HEALTH ANNUAL RESEARCH MEETING SUBSTANCE USE DISORDERS PANEL SESSION JUNE 26, 2017
ACKNOWLEDGEMENTS
Collaborators
Frank Wharam, MD
Marc Larochelle, MD, MPH
Fang Zhang, PhD
Michelle Mello, JD, PhD
Alan Zaslavsky, PhD
2
Funders
Department of Population Medicine, Harvard Medical School & Harvard Pilgrim Healthcare Institute
Harvard University, Graduate School of Arts & Sciences
National Institute of Mental Health
OVERVIEW
Brief background to the opioid epidemic Prescription drug monitoring programs & effectiveness
3
OPIOID SALES, DEATHS, AND TREATMENT ADMISSIONS BY MAJOR DRUG TYPE, UNITED STATES, 1999–2010
4
1 2 3 4 5 6 7 8 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Rate Year
Opioid Sales KG/10,000 Opioid Deaths/100,000 Opioid Treatment Admissions/10,000
Source: CDC. Vital Signs – Overdoses of Prescription Opioid Pain Relievers – United States, 1999-2008. Morbidity and Mortality Weekly Report 2011; 60(43):1487-1492. Updated with 2009 mortality and 2010 treatment admission data.
OPIOID-RELATED OVERDOSE DEATHS CONTINUE TO RISE 2015-16
5
Source: Katz J. Drug Deaths in America are Rising Faster Than Ever. New York Times: The Upshot. June 5, 2017.
OVERVIEW
Brief background to the opioid epidemic Prescription drug monitoring programs & effectiveness
6
What is a PDMP? Electronic systems that digitally store, monitor, & analyze controlled substance dispensing information Who implements PDMPs? 49 States Missouri is only state without one What data do PDMPs collect? Patient info Prescriber info Dispenser info Schedule II-IV drugs Who can access PDMP data? Prescribers Pharmacies Law enforcement State medical boards
7
PDMP EFFECTIVENESS
Mixed older evidence of PDMP impact on drug prescribing & health Newer evidence comparing PDMP features (e.g., mandates)
Use mandates appear to reduce opioid prescribing, measures of doctor
shopping, and overdoses in adults and elderly (esp. with pain clinic laws)
Registration mandates reduce opioid prescribing in Medicaid population
Little known about potential unintended consequences of PDMPs Low/variable prescriber use of PDMPs, but on the rise
8
9
Research Question Hypothesis
10
Administration Operated by health agency Reporting & Monitoring Schedule II-IV Drugs Data collection frequency at least weekly Data Access Prescribers as authorized users Unsolicited reports sent to prescribers Mandatory or automated prescriber enrollment Use of a delegate by prescribers allowed to check PDMP Mandatory consultation of PDMP when prescribing opioids Consultation required under comprehensive circumstances No prescriber immunity for failure to check PDMP
STUDY METHODS
Optum Data: national commercial claims dataset
Designs: comparative interrupted time series with segmented regression; difference-in-differences (DID)
“Exposure” states: 4 states with robust PDMPs
“Comparison” states: 4 states without robust PDMPs
Statistical Analysis:
Marginal effects methods to control for changes in cohort characteristics during the course of study and between state sets calculate adjusted outcome rates; segmented regression analysis with interrupted time series display (quarterly)
Adjusted DID
Main Outcomes:
Percentage of enrollees filling opioid prescriptions (measure of proportion of population receiving)
Mean MEDs per enrollee (measure of strength of opioids prescribed/dispensed)
12
13
14
15
Comparator States (Intervention vs. Control) Variable Difference Preimplementation, % Policy Effect, % (Δ from Pre-to- Postimplementation)
Level 1.24***
Trend —
Level 0.28*
Trend
—
Level 1.36***
Trend 0.04* 0.12**
Level
Trend — 0.04*
† p<0.1 ** p<0.01
* p<0.05 *** p<0.001
16
17
18
Comparator States (Intervention vs. Control) Variable Difference Pre- implementation (MED) Policy Effect (MED) (Δ from Pre-to- Postimplementation)
Level 26.07***
Trend 2.29***
Level 77.86*** 10.08*** Trend 1.29
Level 67.89***
Trend 1.44*** 1.73†
Level
Trend — —
† p<0.1 ** p<0.01
* p<0.05 *** p<0.001
DID RESULTS: THOSE WITH OPIOID RECEIPT
19
Outcome Per Enrollee Mean Change From Baseline to Follow-up, Exposure vs Comparison Exposure Comparison Absolute Change Relative Change, % Pre Post Pre Post Est (95% CI) Est (95% CI) KY vs. MO (n=55,654) Opioid Fills (no.) 2.37 2.12 2.11 2.26
(-0.46,-0.32)***
(-18.71,-13.60)*** Mean MEDs 4177 4025 3922 4628
(-23.53,-13.13)*** NM vs. TX (n=173,860) Opioid Fills (no.) 2.35 2.34 1.88 2.00
(-0.22,-0.05)***
(-10.16,-3.42)*** Mean MEDs 3731 4409 7395 7802
(-861,-320)†
(-17.83,-3.62)** TN vs. GA (n=65,623) Opioid Fills (no.) 2.24 2.19 1.97 2.04
(-0.17,-0.06)***
(-7.81,-2.79)*** Mean MEDs 5898 6019 4072 4370
(-851,-43)*
(-16.93,-3.93)**
DID RESULTS: THOSE WITH OPIOID RECEIPT MEASURES OF “HIGH” OPIOID USE KY VS. MO
20
Outcome Mean Change From Baseline to Follow-up, Exposure vs Comparison Exposure Comparison Absolute Change Relative Change, % Pre Post Pre Post Est (95% CI) Est (95% CI) KY vs. MO (n=55,654) Enrollees w/ Daily MED ≥ 100 (%) 0.94 0.83 0.78 0.87
(-0.32,-0.07)*
(-32.03,-8.80)** Mean Quarters Opioid Rx filled w/ ≥ 3 Drs. Per enrollee 0.04 0.03 0.03 0.04
(-0.02,-0.02)***
(-50.36,-30.54)*** Mean Quarters Opioid Rx filled w/ ≥ 3 Pharmacies per enrollee 0.02 0.01 0.02 0.02
(-0.01,-0.00)***
(-52.72,-23.39)***
CONCLUSIONS
Modest ↓ in proportion filling opioid Rx’s ↓ in MEDs dispensed per enrollee
NY: mandate exceptions, prescribing levels about ½ that of other
exposure states.
21
CONCLUSIONS (CONT’D)
Strongest PDMPS do seem to be having an effect on opioid prescribing and
use (MEDs, fills, high opioid use measures in KY)
Robust policies affect MEDs the most
KY: may exemplify a “model” PDMP Key program features include:
PDMP registration & comprehensive use mandates, delegate access Increased administrative staffing to support PDMP operations Working with prescribers to tweak aspects of PDMP/make user-friendly ASAP: interstate data sharing & integration with medical records (“push” system)
22
STUDY LIMITATIONS
Lack of payments made outside of insurance visible in data Not a lot of post-period for some policies (e.g., NY – 5 quarters) Assumption that state of residence is where drugs are prescribed/filled Focus on opioid prescribing and use, rather than opioid-related injuries Did not explicitly account for cointerventions at state, provider, insurer level.
Many policies in prescribing space, but most significantly weaker than ones studied. Results consistent & tied closely to intervention timeframe across strongest programs.
No distinguishing between appropriate and inappropriate prescribing.
23
24
25
PRESCRIPTION DRUG DUAL PUBLIC HEALTH CHALLENGES DUAL PUBLIC HEALTH OPIOID CHALLENGES
26
Opioid Misuse and Overdose Under-Treatment
SOURCES OF PAINKILLERS
27
Source: NSDUH 2013. http://www.samhsa.gov/data/sites/default/files/NSDUHresultsPDFWHTML2013/Web/NSDUHresults2013.htm
28
OPIOID MISUSE & ABUSE PREVENTION MEASURES
Source: Strang J et al. Drug policy and the public good: evidence for effective interventions. Lancet 2012;379(9810):71-83.
MAT Naloxone PDMP Urine testing Rx limits Rx education Drug approval
Drug importation controls
29
POLICY OPTIONS TO ADDRESS RX DRUG ABUSE
Federal Government: ONDCP, Surgeon General/HHS, SAMHSA, CDC, FDA, DEA State/Local Government & Community: PDMPs, Rx drug laws, prescriber guidelines, needle exchanges pill mill crackdowns, consumer lawsuits Insurer/PBMs: claims analysis, formularies, monitoring, reimbursement incentives Providers: urine testing, pain contracts, prescriber education/guidelines, screening tools, feedback, care integration Drug Manufacturers: formulations, prescriber education
PDMP “BEST PRACTICES” – SOMEWHAT EVIDENCE-BASED
30
Source: PEW Charitable Trusts & Institute for Behavioral Health, Heller School for Social Policy & Management, Brandeis University, Dec. 2016
PDMP PRESCRIBER USE MANDATES AND THEIR POTENTIAL EFFECTS State (Year) Reported Effect on PDMP Queries Reported Effect on Prescribing
Kentucky (2012)
2011 to 2.7 million in 2012 and 4.6 million in 2013
doses in the year before to 6.8 million in the year after the mandate T ennessee (2013)
124,000/month in 2011 to 415,000/month in 2013
July 2013 New York (2013)
11,000/month in the 3.5 years pre- to 42,300/day in the 6 months post-mandate
2013
31
32
PDMP Laws Robust Data Access (Including Mandates) Comprehensive Reporting/ Monitoring Longer Operation under a Health Agency Reduced Lower-Risk, Higher Benefit Rx Increased Lower-Risk, Higher Benefit Rx Reduced Higher-Risk, Lower Benefit Rx Compromised Pain Management Facilitated Pain Management Reduced Misuse
Intervention PDMP Policy Levers Prescriber Behavior Health Outcomes
Increased PDMP Use Increased Prescriber Awareness re: Opioid Rx Risks
Reduced Hospit- alizations Reduced ED Visits
Increased ED Visits / hospitaliz- ations
Prescriber PDMP awareness high; PDMP use variable, but increasing
*2014 national survey of PCPs: 72% aware of state’s PDMP; 53% used (Rutkow et al., H Affairs 2015)
2012 survey of ED residents & attendings: 96% & 81% aware of PDMP; 79% & 51% used PDMP (Feldman et al. Pain Med 2012)
2011 OH survey of Drs. (5 specialties): 84% aware of PDMP; 59% used (Feldman et al. J Pain & Palliative Care Pharmacother 2011)
2003 VA survey of Drs.: 60% aware of PDMP; 11% used (Barrett & Watson J Pain & Palliative Care Pharmacother 2005)
PDMPs do impact prescribing behavior
*OH study of ED Drs.: 61% of patients received fewer/no opioids; 39% of patients received more pain prescriptions (Baehren et al. Ann Int Med 2010)
MA study of ED Drs.: 3.5% of patients did not receive Rx originally planned; 6.5% of patients received Rx not originally planned (Weiner et al. Ann Emerg Med 2013)
FL study of ED Drs.: No change in avg. # of controlled substances prescribed, but Drs. felt data altered their prescribing and were more contented prescribing controlled substances (McAllister et al. Am J Emerg Med 2015)
33
POLICY IMPLICATIONS
Robust PDMPs are a promising intervention to reduce opioid prescribing, including
that which is potentially high risk.
To maximize PDMP utility, must:
Balance various goals and stakeholder interests (patients, providers, law enforcement) Incorporate growing evidence base into policy Provide adequate funding for programs
PDMPs not a panacea and should be evaluated/used in combination with other Rx
drug abuse interventions.
34