Reversing the Tide: Tracking and Responding to Opioid Misuse State - PowerPoint PPT Presentation

Reversing the Tide: Tracking and Responding to Opioid Misuse State Mandatory Review of PDMP by Providers and Pain Clinic Law Reduces ED Visits and Inpatient Stays for Opioid Overdose Kun Zhang, Ph.D. Health Scientist 06/27/2017 Annual Research Meeting National Center for Injury Prevention and Control Division of Unintentional Injury Prevention

Collaborators  Kun Zhang, PhD, Health Scientist, Division of Unintentional Injury Prevention (DUIP), NCIPC, CDC  Gery Guy, PhD, MPH, Senior Health Economist, DUIP, NCIPC, CDC  Rita Noonan, PhD, Branch Chief, DUIP, NCIPC, CDC  Chris Jones, PhD, PharmD, MPH, Director, Division of Science Policy, ASPE, HHS

Disclaimer The results and conclusions in this study are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention and HHS.

Overview  Background and Literature Review  Study Objective  Methods  Results  Conclusions  Limitations

Overprescribing of Opioids

Overprescribing of Opioids

Opioids Driving the Epidemic of Overdose Deaths 16 All Drug Overdoses Deaths per 100,000 population 14 12 10 Opioid Overdoses 8 (including Rx and heroin) 6 4 2 0 2000 2002 2004 2006 2008 2010 2012 2014 SOURCE: CDC/NCHS National Vital Statistics System Mortality File.

Opioids-related ED Visits and Inpatient Stays

Promising State Policies to Address the Epidemic by Improving Prescribing  Prescription Drug Monitoring Program (PDMP) – State electronic database that tracks controlled substances dispensing ‒ Implemented in 49 states and D.C. ‒ DEA schedules II-IV such as opioid, benzodiazepine, stimulants, etc. ‒ Collect information of the filled prescription – PDMPs can help ‒ Identify patients at risk for opioid abuse and overdose ‒ Inform providers of other medications patient receiving that may interact with those prescribed  Pain clinic laws (pill mill laws) – State law that regulates pain clinics on – Ownership, licensure, inspection, limited days supply, etc .

Policy Definition in the Study  Mandatory PDMP review – All prescribers with DEA number required to register with PDMP – All prescribers with DEA number required to check PDMP data prior to prescribing opioids – 2011: OH; 2012: KY, WV; 2013: TN, NY; 2014: NM; 2015: NV, OH; 2016: CT, MA, OK, NJ; 2017: CA, WI  Pain clinic laws – Registration, certification, licensing, physician on-site, medical director required, prescribing restrictions, penalties, etc. – 2007: LA; 2009: TX; 2010: FL; 2011: OH; 2012: KY, TN; 2013: WV

Literature Review  Mandatory PDMP review – Reduced overall prescribing of opioids PDMP Center of Excellence, Brandeis University (2014) Kentucky House Bill 1 Impact Evaluation, University of Kentucky (2015) Rasubala L, Pernapati L, Velasquez X, et al. (2015), Wen H, Schackman BR, Aden B, Bao Y (2017) Dowell D, Zhang K, Noonan R, Hockenberry J (2016) – Reduced inappropriate prescribing of opioids PDMP Center of Excellence, Brandeis University (2014) Kentucky House Bill 1 Impact Evaluation (2015) Buchmueller TC, Carey C (2017) – Reduced prescription opioid overdose deaths Dowell D, Zhang K, Noonan R, Hockenberry J (2016)  Pain clinic laws – Reduced overall prescribing of opioids Rutkow L, Chang H, Daubresse M, et al. (2015), Lyapustina T, Rutkow L, Chang H, et al. (2016) – Reduced diversion of prescription opioids Surratt HL, O’Grady C, Kurtz SP, et al. (2014) – Reduced prescription opioid overdose deaths Hendricks A, Richey M, McGinty EE, et al. (2015)

Objective and Selection of State  Study the impact of mandatory PDMP review and pill mill law on opioid overdose related ED visits and inpatient stays  Select the states of study – Mandatory review and pill mill law: state-level policy – State-level data availability – Literature and institutional knowledge – Intervention state: Kentucky – Control state: North Carolina

Policies in Kentucky  Mandatory review of PDMP: Effective Drugs State date Policy detail included Timing triggers Exceptions End of life care; cancer pain; inpatient at hospital or long-term Mandatory review of care facility; during an DEA Initial prescription, Kentucky July, 2012 PDMP by prescribers emergency or following surgery; Schedules II-IV every 90 days prior to prescribing single-dose treatments to relieve symptoms from a procedure.  Pain clinic law: registration, certification, licensing, physician on-site, medical director required, prescribing restrictions, penalties, etc.

Percent of Active Kentucky Prescribers Registered in PDMP 100 90 80 Implementation of mandate in Kentucky 70 60 50 40 30 20

Reports Requested by Kentucky Prescribers: per 1,000 Residents 300 250 200 Implementation of mandate in Kentucky 150 100 50 0

Data  Data source – HCUP o State Inpatient Databases (SID), State Emergency Department Databases (SEDD) o All payers, entire sample of discharges from community-based hospitals in a state – States: KY, NC – Years: 2011-2013  Drug overdose definitions using ICD9-CM – Prescription opioid overdose o 965.00, 965.02, 965.09, E850.1, E850.2 – Heroin overdose o 965.01, E850.0 – Exclusion o Deceased

Analytical Approach  Outcome variable – Overdose visits aggregated at monthly level: prescription opioids, heroin – Converted to rate per 100,000 population  Interrupted time series analysis (ITSA) – Intervention state (KY) and control state (NC): comparative ITSA – 18 pre-intervention time points, 17 post-intervention time points – Nonequivalent dependent variable: other type of overdose (cocaine overdose)  Model specification 𝒁 𝒖𝒌 = 𝜸 𝟏 + 𝜸 𝟐 (𝒏𝒑𝒐𝒖𝒊 𝒖𝒌 + 𝜸 𝟑 𝒒𝒑𝒎𝒋𝒅𝒛 𝒖𝒌 + 𝜸 𝟒 𝒒𝒑𝒎𝒋𝒅𝒛 × 𝒏𝒑𝒐𝒖𝒊 𝒖𝒌 + 𝜸 𝟓 𝒕𝒖𝒃𝒖𝒇 𝒖𝒌 ൯ + 𝜸 𝟔 𝒕𝒖𝒃𝒖𝒇 × 𝒏𝒑𝒐𝒖𝒊 𝒖𝒌 + 𝜸 𝟕 𝒕𝒖𝒃𝒖𝒇 × 𝒒𝒑𝒎𝒋𝒅𝒛 𝒖𝒌 + 𝜸 𝟖 𝒕𝒖𝒃𝒖𝒇 × 𝒒𝒑𝒎𝒋𝒅𝒛 × 𝒏𝒑𝒐𝒖𝒊 𝒖𝒌 + 𝜻 𝒖𝒌

Results: Prescription Opioid Overdose Kentucky and average of controls Intervention starts: 630 • 8 Same pre-policy trend • Level change 7 immediately following the intervention: – 1.16** 6 • Pre and post intervention slope difference: – 0.09* 5 • Monthly change post intervention: – 0.08*** 4 *** P <0.01 ** P <0.05 * P <0.1 2011m1 2012m1 2013m1 2014m1 Month Kentucky: Actual Predicted Controls average: Actual Predicted Prais-Winsten and Cochrane-Orcutt regression - lag(1)

Results: Heroin Overdose Kentucky and average of controls Intervention starts: 630 • 5 Different pre-policy trend • 4 Level change immediately following the intervention: 0.95** 3 • Pre and post intervention 2 slope difference: – 0.07* 1 • Monthly change post intervention: 0.027 0 2011m1 2012m1 2013m1 2014m1 *** P <0.01 ** P <0.05 * P <0.1 Month Kentucky: Actual Predicted Controls average: Actual Predicted Prais-Winsten and Cochrane-Orcutt regression - lag(1)

Results: Benzodiazepine Overdose Kentucky and average of controls Intervention starts: 630 • 10 Same pre-policy trend • Level change 9 immediately following the intervention: – 1.46** 8 • Pre and post intervention 7 slope difference: – 0.09 6 • Monthly change post intervention: – 0.1** 5 2011m1 2012m1 2013m1 2014m1 *** P <0.01 ** P <0.05 * P <0.1 Month Kentucky: Actual Predicted Controls average: Actual Predicted Prais-Winsten and Cochrane-Orcutt regression - lag(1)

Results: Cocaine Overdose Kentucky and average of controls Intervention starts: 630 2 • Same pre-policy trend 1.5 • Nothing significant 1 *** P <0.01 ** P <0.05 * P <0.1 .5 0 2011m1 2012m1 2013m1 2014m1 Month Kentucky: Actual Predicted Controls average: Actual Predicted Prais-Winsten and Cochrane-Orcutt regression - lag(1)

Conclusions  Mandatory PDMP review + pain clinic laws – Decreased ED visits and inpatient stays of prescription opioid overdose o Both level change and slope change o Level change is immediate; sustained decrease in the study period – Mixed results for ED visits and inpatient stays of heroin overdose o Violation of ITSA assumption, significant increasing trend pre-policy in Kentucky o Immediate level increase, but reversed the trend of increasing – No impact on cocaine overdose  Consistent with previous findings – Decreased overall opioid prescribing and prescription opioid overdose deaths – Associated with decrease in high-risk prescribing of opioids  Promising interventions for states to consider, particularly those with low PDMP use, high rates of risk indicators for prescribing, and high rates of opioid overdose

Limitations  Cannot differentiate the effect of mandatory PDMP review from the effect of pain clinic laws – Implemented simultaneously – No data on pain clinics – Pain clinic law alone insufficient to impact overall prescribing  The regression method used assumes linear trends over time  Single comparison group

THANK YOU Contact Information: Kun Zhang: kzhang@cdc.gov