Prescription Drug Monitoring Programs as a Key Tool in Addressing - - PowerPoint PPT Presentation

▶
prescription drug monitoring programs as a key tool in
SMART_READER_LITE
LIVE PREVIEW

Prescription Drug Monitoring Programs as a Key Tool in Addressing - - PowerPoint PPT Presentation

Nov 10, 2023 117 likes •463 views

Institute for Behavioral Health SCHNEIDER INSTITUTES FOR HEALTH POLICY Prescription Drug Monitoring Programs as a Key Tool in Addressing the Opioid Crisis Cindy Parks Thomas and Peter N. Kreiner Academy Health Annual Research Meeting June 27,

slide-1
SLIDE 1

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Prescription Drug Monitoring Programs as a Key Tool in Addressing the Opioid Crisis

Cindy Parks Thomas and Peter N. Kreiner Academy Health Annual Research Meeting June 27, 2017

slide-2
SLIDE 2

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Overview

  • Background: Prescription Monitoring Programs

– Data and uses

  • Prescription Behavior Surveillance System (PBSS)
  • Types of epidemiological studies conducted

– Prescription Behavior Surveillance System (PBSS) data trends and comparison studies – Studies to assess PBSS data quality and consistency – Evaluations of PDMP law, policy, and practice – Studies using PDMP data linked with other health data

2

slide-3
SLIDE 3

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Background: Prescription Drug Monitoring Programs

  • 49 states, all retail controlled substance prescriptions

reported to state agency; 32 states require checks

  • Data

– Prescription: date, medication, dose, quantity, days supply – Prescriber, pharmacy, location, patient age, gender, ID, CS prescription history

  • Uses

– Prescribers and dispensers check for concurrent prescriptions or prescribers – Identify patterns and questionable activity – high prescribers, multiple prescribers, prescriptions pharmacies – Analytic: epidemiologic studies and surveillance – early warning or growth patterns by geographic area

3

slide-4
SLIDE 4

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Background: Prescription Behavior Surveillance System (PBSS)

  • Funded by CDC and FDA
  • Longitudinal, multi-state database of de-identified PDMP

data, to serve as:

– Early warning public health surveillance tool – Policy and law evaluation tool

  • Currently houses data from 12 states: CA, DE, FL, ID, KY, LA,

ME, OH, TX, VA, WA, WV

  • Data mostly from 2010 to present, updated quarterly
  • 43 descriptive measures and indicators of patient, prescriber,

and pharmacy risk behaviors

– compiled quarterly for each state

4

slide-5
SLIDE 5

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Prescription Behavior Surveillance System (PBSS) metrics examples

  • Overall usage within drug classes and for selected individual drugs
  • Daily dosage
  • Overlapping prescriptions within each drug class, across the opioid and

benzodiazepine classes, and across dosage forms of opioid analgesics (i.e., immediate vs. extended release)

  • Questionable activity within a class or classes
  • Payment sources
  • Pharmacy-based measures of possible inappropriate dispensing

Each participating PDMP provides an initial set of legacy data, with subsequent quarterly updates, enabling an ongoing surveillance of prescription activity trends.

5

slide-6
SLIDE 6

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

PBSS-Related Projects and Studies I

  • PBSS data trends and comparisons

– User-friendly state data briefs and multiple state issue briefs – Comparison of prescribing measures across states, 2013

(Paulozzi, Strickler, Kreiner, & Koris, 2015)

– Comparison of prescribing measure and risk indicator trends across states – Examination of veterans’ prescriptions and risk indicators when using community providers vs. VHA providers (Becker

et al., 2017)

– Examination of prescription dosages by physician specialty in Ohio (Weiner et al., 2017)

6

slide-7
SLIDE 7

0.00 50.00 100.00 150.00 200.00 250.00 Jan- Mar 2010 Apr- Jun 2010 Jul- Sep 2010 Oct- Dec 2010 Jan- Mar 2011 Apr- Jun 2011 Jul- Sep 2011 Oct- Dec 2011 Jan- Mar 2012 Apr- Jun 2012 Jul- Sep 2012 Oct- Dec 2012 Jan- Mar 2013 Apr- Jun 2013 Jul- Sep 2013 Oct- Dec 2013 Jan- Mar 2014 Apr- Jun 2014 Jul- Sep 2014 Oct- Dec 2014 Jan- Mar 2015 Apr- Jun 2015 Jul- Sep 2015 Oct- Dec 2015 Jan- Mar 2016 Apr- Jun 2016 Jul- Sep 2016 Oct- Dec 2016

Hydrocodone SA Prescriptions per 1,000 State Residents By Quarter

CA DE FL ID KY LA ME OH TX VA WV

Hydrocodone rescheduled

7

slide-8
SLIDE 8

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

PBSS-Related Projects and Studies II

  • Understanding and assessing PDMP data

– Methods to assess data quality

  • Missing data, out-of-bounds values and their effects on

surveillance measures

– Documentation and assessment of methods to determine which prescriptions belong to the same patient (record linking) – Effects of different record-linking methods on descriptive measures and risk indicators

8

slide-9
SLIDE 9

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

PDMP Record-Linking Comparison: California

Measure (2013) Exact match only Probabilistic match Percent change MPE count 718 3920 446.0% Average daily opioid dosage (MME) 48.59 48.70 < 1 % Percent with > 90 MME 12.17 12.12 < 1% Percent overlapping

  • pioid days

17.09 19.69 15.2% Percent overlapping

  • pioid and

benzodiazepine days 11.93 13.56 13.7% Percent overlapping LA and SA opioid days 8.61 9.27 7.7%

9

slide-10
SLIDE 10

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

PBSS-Related Projects and Studies III

  • Evaluations of PDMP law, policy, and practice

– Evaluation of pilot electronic prescribing of controlled substances project, Berkshire Health System (Thomas et al.,

2011, 2013)

– Evaluation of Massachusetts PDMP unsolicited reporting

(Thomas et al., 2014; Young, Kreiner, & Panas, 2017)

– Effects of state mandatory prescriber use of PDMP laws on PDMP usage, prescribing measures, and risk indicators

(Kreiner, 2017)

10

slide-11
SLIDE 11

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Reports Requested by In-State Prescribers: Rate per 1,000 Residents

Footnote: Indiana’s data is not available after Q1 of 2015. 11

slide-12
SLIDE 12

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Overlapping (7 Days) Opioid Prescriptions: Rate per 1,000 Residents

Footnote: Indiana’s data is not available after Q1 of 2015. 12

slide-13
SLIDE 13

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

PBSS-Related Projects and Studies IV

  • PDMP data linked with other health data

– Changes in rates of heroin-related vs. prescription opioid- related death rates in Kentucky counties – Buprenorphine prescribing by waivered physicians (Thomas

et al., under review)

– Validation study of prescriber risk indicators (Kreiner et al.,

2017)

– Validation study of patient risk indicators

  • Patient behavior trajectory analysis

– Effects of medical board sanctions on subsequent prescriber behavior and effects on their patients

13

slide-14
SLIDE 14

Prescription Opioid-related Death Rates per 100,000 Population by County, 2011-2012 and 2013-2014

Prescription opioid-related death rates 0 - 1.57 1.58 - 2.29 2.30 - 2.94 2.95 - 4.39 4.40 - 7.86 Prescription opioid-related death rates 0 - 2.00 2.01 - 2.63 2.64 - 3.27 3.28 - 3.97 3.98 - 7.24

14

slide-15
SLIDE 15

Heroin-related Death Rates per 100,000 Population by County, 2011-2012 and 2013-2014

Heroin-related death rates 0.01 - 1.42 1.43 - 1.69 1.70 - 2.43 2.44 - 5.66

Heroin-related death rates 0.01 - 1.51 1.52 - 1.98 1.99 - 2.65 2.66 - 4.65

15

slide-16
SLIDE 16

Kentucky: Opioid Prescriptions per Resident 2011-2012 Average, by County

Opioid prescriptions per resident 0.84 - 1.05 1.05 - 1.15 1.15 - 1.25 1.25 - 1.43 1.43 - 1.85

16

slide-17
SLIDE 17

Findings from Spatial Regressions

  • Prescription opioid-

related death rates in 2013-2014 associated with:

– Population density (p < .01) – % Hispanic non-White (p = .03) – % White (p = .02) – Opioid prescription rate (p < .001) – R2 = .563

  • Heroin-related death rates

in 2013-2014 associated with:

– Population density (p < .001) – Population mobility (p = .02) – Percent White (p = .02) – Spatial lag of heroin-related death rate in 2011-2012 (p < .001) – R2 = .400

17

slide-18
SLIDE 18

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Buprenorphine prescribing patterns study

Waivered physician prescribing will likely be critical to addressing insufficient MAT capacity

  • Approximately 2% of U.S. physicians are waivered
  • An estimated 33% of waivered physicians are not

prescribing buprenorphine

  • Several small studies suggest that many prescribers

treat relatively few patients

Funded by ASPE, with RAND Corp.

18

slide-19
SLIDE 19

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Using PBSS: What we wanted to learn

  • How many patients were buprenorphine prescribers

actually treating?

  • What was treatment episode duration (long

episodes could constrain the total number of patients treated)?

  • To what extent might patient limits be constraining

treatment?

19

slide-20
SLIDE 20

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

ASPE study of waivered physicians

20

223,000+ patients

4

YEARS

~10,600 prescribers + waiver status 3 states (CA, OH, ME) 4-year period (2010–2013)

NOTES:

  • All prescribers who prescribed buprenorphine at any point during the study period
  • PDMP data representing 100% of prescriptions filled at retail pharmacies,

excluding those indicated for pain

slide-21
SLIDE 21

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Buprenorphine waivered physician study findings

21

  • Patients treated per month

– Median: 4 – Mean: 13 – Near cap:

  • 30-patient prescribers: 8.8% of months
  • 100-patient prescribers: 10.3% of months
  • Treatment duration

– Median: 31 days – Mean: 95 days – 75th percentile:

  • 30-patient prescribers: 87 days
  • 100-patient prescribers: 103 days
slide-22
SLIDE 22

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Buprenorphine waivered physician study findings, cont.

22

  • Study months in which prescribers had no patients on

buprenorphine:

– 30 patient prescribers: 60.4% – 100-patient prescribers: 14.2%

  • Percent of prescribers who treated no patients with

buprenorphine during the study

– 30 patient prescribers: 42.7% – 100-patient prescribers: 7.3%

  • Percent of prescribers who treated a patient with

buprenorphine the entire study time

– 30 patient prescribers: 18.9% – 100-patient prescribers: 75.8%

slide-23
SLIDE 23

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Patient Behavior Trajectory Studies: Example

  • Patient risk indicators computed monthly:

– Average daily opioid dosage (> 100 MME) – Opioid prescriptions for 90 consecutive days – Overlapping opioid prescriptions (>= 7 days overlap) – Overlapping opioid and benzodiazepine prescriptions (>= 7 days

  • verlap)

– Multiple provider episodes (>= 4 prescribers and 4 pharmacies in 1 month)

  • We selected individuals who exceeded at least one risk threshold for at

least one month in 2014

  • We examined their behavior on these measures for the 36 months Jan.,

2013 – Dec., 2015, SAS proc traj

23

slide-24
SLIDE 24

24

slide-25
SLIDE 25

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

25

slide-26
SLIDE 26

To What Extent Do Individuals Who Exceed One Risk Threshold Exceed Others?

1 Risk Threshold Only 2 Risk Thresholds Only 3 Risk Thresholds Only 4 Risk Thresholds Only 5 Risk Thresholds DE: Total 53,033 patients Percent exceeding threshold(s) 55.3% 20.9% 16.3% 7.4% 0.1% Ohio: Total 671,439 patients Percent exceeding threshold(s) 60.8% 24.1% 11.7% 3.4% 0.1%

26

slide-27
SLIDE 27

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Conclusions from trajectory study

  • For each of the opioid measures, the most prevalent

trajectory group indicates exceeding the risk threshold is a

  • ne-time or infrequent occurrence
  • The different trajectories, consistent across three states,

suggest different risk profiles needing different responses or interventions

– How to take a pattern of behavior over time into account in generating alerts?

  • Future work will examine associations of the trajectory groups

with (1) patient demographics and location, and (2) overdose death outcomes

27

slide-28
SLIDE 28

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Summary

28

  • PBSS prescription monitoring program data powerful

tool for

– surveillance – targeting prevention and treatment – Evaluating policies

  • Merging with other public health datasets expands

potential

  • Additional studies in process and planned
slide-29
SLIDE 29

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Citations

  • PBSS measures:

http://www.pdmpassist.org/pdf/COE_documents/Add_to_TTAC/Definitions% 20of%20PBSS%20Measures.pdf

  • PBSS background and data briefs:

http://www.pdmpassist.org/content/prescription-behavior-surveillance- system

  • PBSS surveillance paper: Paulozzi, LJ, Strickler, GK, Kreiner, PW, and Koris, CM.

Controlled substance prescribing patterns – Prescription Behavior Surveillance System, Eight States, 2013. Morbidity and Mortality Weekly Report, Surveillance Summaries; 64(9): 1-14. October 16, 2015. Available at: https://www.cdc.gov/mmwr/preview/mmwrhtml/ss6409a1.htm

  • Becker, WC, Fenton, BT, Brandt, CA, Doyle, EL, Francis, J, Goulet, JL, Moore,

BA, Torrise, V, Kerns, RD, and Kreiner, PW. (In press). Multiple sources of prescription payment and risky opioid therapy among veterans. Medical Care.

29

slide-30
SLIDE 30

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Citations II

  • Weiner, SG, Baker, O, Rogers, AF, Garner, C, Nelson, LS, Kreiner, PW, and

Schuur, JD. (2017). Opioid prescriptions by specialty in Ohio, 2010-2014. Pain Medicine pnx027; March 6, 2017. doi: 10.1093/pm/pnx027

  • Thomas, C., Kim, M., Kelleher, S. Nikitin, R., Kreiner, P., McDonald, A., and

Carrow, G. (2013). Early experience with electronic prescribing of controlled substances in a community setting. Journal of the American Medical Informatics Association, doi: 10.1136/amiajnl-2012-001499.

  • Thomas, C., Kim, M., McDonald, A., Kreiner, P., Kelleher, S., Blackman, M.,

Kaufman, P., & Carrow, G. (2011). Prescribers’ expectations and barriers to electronic prescribing of controlled substances. Journal of the American Medical Informatics Association, doi:10.1136/amaiamnl-2011-000209.

  • Thomas, C.P., Kim, M., Nikitin, R.V., Kreiner, P., Clark, T., and Carrow, G.

(2014) Physician response to unsolicited prescription drug monitoring program reports in Massachusetts. Pharmacoepidemiology and Drug Safety, 23: 950-957, doi:10.1002/pds.3666.

30

slide-31
SLIDE 31

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Citations III

  • Evaluation of unsolicited reporting in MA: Young, L, Kreiner, P, and Panas,
  • L. (2017). Unsolicited Reporting to Prescribers of Opioid Analgesics by a

State Prescription Drug Monitoring Program: An Observational Study with Matched Comparison Group. Pain Medicine pnx044; April 4, 2017. doi: 10.1093/pm/pnx044

  • Kreiner, PW. PDMP Policy Developments: Evaluating Prescriber-Use
  • Mandates. Presentation at National Rx and Heroin Abuse Summit, Atlanta,
  • GA. April 18, 2017.
  • Initial validation study of PBSS prescriber risk indicators: Kreiner, PW,

Strickler, GK, Undurraga, EA, Torres, ME, Nikitin, RV, and Rogers, A. (2017). Validation of prescriber risk indicators obtained from prescription drug monitoring program data. Drug and Alcohol Dependence, 173: S31-S38. doi: 10.1016/j.drugalcdep.2016.11.020. Available at (open source): http://www.sciencedirect.com/science/article/pii/S0376871616310298

31

slide-32
SLIDE 32

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Spatial Regression I: Heroin-related Death Rates

Dependent variable: heroin-related death rates 2013 - 2014 Independent variable Coefficient

  • Std. Error

z-value Probability Constant .634 4.930 .129 .898 Population density .064 .019 3.314 <.001 Population mobility 1.041 .441 2.361 .018 Poverty rate

  • .074

.141

  • .527

.598 Percent Hispanic ,215 .241 .894 .372 Percent White .943 .418 2.26 .024 Prescription opioid- related death rate 2011

  • 2012

.033 .049 .680 .496 Spatially-lagged version of heroin- related death rate 2011

  • 2012

.922 .144 6.400 <.001 Spatial error term .066 .138 .476 .634 R2: .563 Diagnostics for heteroscedasticity Breusch-Pagan Test DF: 7 Value: 6.395 Probability: .494 Diagnostics for spatial dependence Spatial error likelihood ratio test DF: 1 Value: .191 Probability: .662

slide-33
SLIDE 33

Institute for Behavioral Health

SCHNEIDER INSTITUTES FOR HEALTH POLICY

Spatial Regression II: Prescription Opioid-related Death Rates

Dependent variable: prescription opioid-related death rates 2013 - 2014 Independent variable Coefficient

  • Std. Error

z-value Probability Constant

  • 15.406

5.636

  • 2.734

.006 Population density .075 .026 2.835 .005 Population mobility

  • .094

.164

  • .573

.567 Poverty rate

  • .029

.205

  • .142

.887 Percent Hispanic .725 .339 2.140 .032 Percent White 1.225 .536 2.288 .022 Opioid prescription rate 4.693 .816 5.753 <.001 Heroin-related death rate 2011 - 2012

  • .134

.120

  • 1.114

.265 Spatially-lagged version of prescription

  • pioid-related death

rate 2011 - 2012 .059 .121 .487 .627 Spatial error term .042 .140 .298 .766 R2: .400 Diagnostics for heteroscedasticity Breusch-Pagan Test DF: 7 Value: 15.480 Probability: .030 Diagnostics for spatial dependence Spatial error likelihood ratio test DF: 1 Value: .209 Probability: .648