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 ANNUAL RESEARCH MEETING SUBSTANCE USE DISORDERS PANEL SESSION JUNE 26, 2017
ACKNOWLEDGEMENTS Collaborators Funders Frank Wharam, MD Department of Population Medicine, Harvard Medical School & Harvard Marc Larochelle, MD, MPH Pilgrim Healthcare Institute Fang Zhang, PhD Harvard University, Graduate School of Michelle Mello, JD, PhD Arts & Sciences Alan Zaslavsky, PhD National Institute of Mental Health 2
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 Opioid Sales KG/10,000 Opioid Deaths/100,000 Opioid Treatment Admissions/10,000 8 7 6 5 Rate 4 3 2 1 0 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Year 4 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
Prescription Drug Monitoring Programs (PDMPs) What data do Who can access Who implements What is a PDMP? PDMPs collect? PDMP data? PDMPs? Patient info Electronic systems 49 States Prescribers that digitally store, Prescriber info monitor, & analyze Missouri is only Pharmacies Dispenser info controlled state without one Law enforcement substance Schedule II-IV dispensing drugs State medical information 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
Research Question Hypothesis 1. Do robust PDMPs decrease the percentage of Yes enrollees obtaining an opioid Rx? or the number of opioid fills? 2. Do robust PDMPs decrease the strength of Yes opioids (MEDs) dispensed per enrollee? 3. Do robust PDMPs decrease measures of Yes potentially high-risk opioid use among those with prescription opioid receipt? 9
Category Robust PDMP Features 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 10 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)
INTERVENTION (PINK) & CONTROL STATES (BLUE) 12
PERCENTAGE OF ENROLLEES FILLING OPIOID RX’S PER QUARTER 13
PERCENTAGE OF ENROLLEES FILLING OPIOID RX’S PER QUARTER 14
Effect of Robust PDMPs on % Enrollees Filling Opioid Rx / Quarter Comparator States Variable Difference Policy Effect, % (Intervention vs. Control) Preimplementation, ( Δ from Pre-to- % Postimplementation) a. KY vs. MO Level 1.24*** -1.30*** Trend — -0.03 b. NM vs. TX Level 0.28* -0.40 Trend -0.03 — c. TN vs. GA Level 1.36*** -0.81** 0.04 * Trend 0.12** d. NY vs. NJ Level -0.65*** -0.33*** Trend — 0.04* † p<0.1 ** p<0.01 15 * p<0.05 *** p<0.001
MEAN MED PER ENROLLEE PER QUARTER 16
MEAN MED PER ENROLLEE PER QUARTER 17
Effect of Robust PDMPs on Mean MED Dispensed / Enrollee / Quarter Comparator States Variable Difference Pre- Policy Effect (MED ) (Intervention vs. Control) implementation ( Δ from Pre-to- (MED) Postimplementation) a. KY vs. MO Level 26.07*** -31.60*** Trend 2.29*** -5.06*** b. NM vs. TX Level 77.86*** 10.08*** Trend 1.29 -7.44*** 67.89 *** c. TN vs. GA Level -27.14*** 1.44 *** 1.73 † Trend d. NY vs. NJ Level -26.53*** -5.57** Trend — — † p<0.1 ** p<0.01 18 * p<0.05 *** p<0.001
DID RESULTS: THOSE WITH OPIOID RECEIPT NO. OF FILLS & MEAN MED Outcome Mean Change From Baseline to Follow-up, Per Enrollee 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) (-0.46,-0.32) *** Opioid Fills (no.) 2.37 2.12 2.11 2.26 -0.39 -16.15 (-18.71,-13.60)*** -858 (-1143,-571.3) *** Mean MEDs 4177 4025 3922 4628 -18.33 (-23.53,-13.13)*** NM vs. TX (n=173,860) (-0.22,-0.05) *** Opioid Fills (no.) 2.35 2.34 1.88 2.00 -0.14 -6.79 (-10.16,-3.42)*** Mean MEDs 3731 4409 7395 7802 -270 (-861,-320) † -10.72 (-17.83,-3.62)** TN vs. GA (n=65,623) Opioid Fills (no.) 2.24 2.19 1.97 2.04 -0.11 (-0.17,-0.06)*** -5.23 (-7.81,-2.79)*** 19 Mean MEDs 5898 6019 4072 4370 -447 (-851,-43)* -10.43 (-16.93,-3.93)**
DID RESULTS: THOSE WITH OPIOID RECEIPT MEASURES OF “HIGH” OPIOID USE KY VS. MO Mean Change From Baseline to Follow-up, Outcome 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 ≥ (-0.32,-0.07) * 0.94 0.83 0.78 0.87 -0.20 -20.42 (-32.03,-8.80)** 100 (%) Mean Quarters Opioid Rx (-0.02,-0.02) *** filled w/ ≥ 3 Drs. Per 0.04 0.03 0.03 0.04 -0.02 -40.44 (-50.36,-30.54)*** enrollee Mean Quarters Opioid Rx 20 filled w/ ≥ 3 Pharmacies 0.02 0.01 0.02 0.02 -0.01 (-0.01,-0.00)*** -38.06 (-52.72,-23.39)*** per enrollee
CONCLUSIONS Robust PDMPs associated with: Modest ↓ in proportion filling opioid Rx’s ↓ in MEDs dispensed per enrollee When use mandate less broad and not paired with registration mandate, effects not as significant: 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
THANK YOU! Questions welcome! haffajee@umich.edu 24
APPENDIX SLIDES 25
PRESCRIPTION DRUG DUAL PUBLIC HEALTH CHALLENGES DUAL PUBLIC HEALTH OPIOID CHALLENGES Under-Treatment Opioid Misuse and of Pain Overdose 26
SOURCES OF PAINKILLERS 27 Source: NSDUH 2013. http://www.samhsa.gov/data/sites/default/files/NSDUHresultsPDFWHTML2013/Web/NSDUHresults2013.htm
OPIOID MISUSE & ABUSE PREVENTION MEASURES Naloxone MAT Urine PDMP testing Rx Rx education limits Drug Drug importation approval controls 28 Source: Strang J et al. Drug policy and the public good: evidence for effective interventions. Lancet 2012;379(9810):71-83.
Recommend
More recommend
