Identifying Barriers to Uptake of Medication Assisted Therapy Among - - PowerPoint PPT Presentation

1

Identifying Barriers to Uptake of Medication Assisted Therapy Among Patients Diagnosed with Opioid Use Disorder

June 4, 2019 Ezra Fishman, Geetanjoli Banerjee, John Barron, and Gosia Sylwestrzak HealthCore, Inc., Wilmington, DE efishman@healthcore.com

2

Disclosures

All authors are employees of HealthCore, Inc., a wholly-owned, independently-operated research subsidiary of Anthem, Inc. Anthem Inc. had no role in the conduct of the analysis or the decision to submit the abstract. Results do not necessarily reflect the opinions or policies of Anthem Inc.

3

Motivation: MAT is effective but under-utilized

• Among patients with opioid use disorder (OUD), medication-

assisted treatment (MAT), in conjunction with psychosocial treatment, reduces non-medical opioid use and all-cause and

• verdose mortality (Sordo 2017 BMJ).
• However, MAT is under-utilized (McCarty 2017 Ann Rev Pub

Health, Saloner 2015 JAMA). Reasons include:

• Persistent expectations of abstinence as the proper treatment course
• Lack of prescribers licensed to dispense MAT

–https://www.samhsa.gov/medication-assisted-treatment/training-materials-resources/buprenorphine-waiver

4

MAT uptake among patients diagnosed with OUD

OBJECTIVE 1 Quantify MAT uptake among patients newly diagnosed with OUD OBJECTIVE 2 Identify barriers to MAT uptake among such patients

5

Study population

• N=70,437 patients below age 65 and newly diagnosed with OUD

between 01 January 2013 and 30 April 2018.

• All patients had ≥12 months’ continuous enrollment in

commercial health plan prior to first OUD diagnosis date (index date).

• We followed patients from the first OUD diagnosis date until the

date of first MAT dispensing/administration (the outcome of interest), death, disenrollment from the health plan, or end of study period.

• Data were queried in August 2018 from our commercial claims

database, the HealthCore Integrated Research Database (HIRDSM).

6

OUD is not only a disorder for young adults

2.5% 16.6% 16.3% 18.1% 23.1% 23.3% 5 10 15 20 25 30 0-17 18-24 25-34 35-44 45-54 55-64

Percent of Cohort (%) Age at Index (years)

7

OUD diagnosis became more common over time

Note: Patients diagnosed in 2018 were excluded from this figure. They represented 3.6% of the cohort because data were only available from the first quarter of 2018.

13.9% 16.8% 20.9% 25.0% 23.5% 5 10 15 20 25 30 2013 2014 2015 2016 2017

Percent of Cohort (%) Index Year

8

Other baseline characteristics of study cohort

89%

OUD diagnosis characterized as moderate to severe (ICD-10 dx code F11.20 or equivalent)

24%

saw a “special specialist” on date of OUD diagnosis

47%

FEMALE

67%

documented Rx fill for

• pioid in baseline period

Rx 29%

fill for benzodiazepine in baseline period

Rx 37%

saw a specialist in psychiatry, addiction medicine, or pain medicine (a “special specialist”) in the baseline period

2.2

mean Elixhauser Comorbidity Index (range: 0-19)

9

16.4% 10.6% 8.1% 10.6% 12.2% 12.6% 10.1% 7.1% 12.3% 2 4 6 8 10 12 14 16 18 20 22 24 26 <1 1-2 3-4 5-7 8-11 12-17 18-23 24-29 30+

Percent of Cohort (%) Months at risk

Distribution of follow-up time

Median: 9.3 mo Mean: 13.4 mo Max: 64 mo

10

Outcome summary

16,577 patients (=24%) initiated MAT in follow-up period 21 patients initiating MAT per 100 person-years

• f follow-up

11

Analytic method

• Cox proportional hazards model
• The following predictors of MAT uptake were included:
• Patient age
• Sex
• Elixhauser comorbidity index (ECI)
• Year of diagnosis of OUD
• Geographic region on date of diagnosis
• Encounter with “special specialist” on index date
• Any benzo Rx in baseline period
• Any opioid Rx fill in baseline period
• Index OUD diagnosis classified as moderate/severe
• Outcome = first MAT dispensing, death, or disenrollment from plan,

whichever comes first

Special specialist = Physician with specialty in psychiatry, addiction medicine, or pain medicine

12

Factors associated with MAT uptake

0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 Adjusted hazard ratio, 95% CI

Male Special Specialist Index Dt Benzo Rx Fill Baseline Opioid Rx Fill Baseline OUD Diagnosis Moderate/Severe

Higher hazard ratio = more likely to initiate MAT

13

0.7 0.8 0.9 1 1.1 1.2 1.3 Adjusted Hazard Ratio, 95% CI

Factors associated with MAT uptake

+1 yr age +1 ECI NE (vs. MW) S (vs. MW) W (vs. MW) Higher hazard ratio = more likely to initiate MAT

14

0.5 0.6 0.7 0.8 0.9 1 1.1 Adjusted hazard Ratio, 95% CI

After adjusting for covariates, each year of diagnosis after 2013 was associated with lower MAT uptake

2014 2015 2016 2017 2018 Higher hazard ratio = more likely to initiate MAT

15

Summary

• Only ¼ of newly diagnosed OUD patients initiated MAT.
• Estimate from National Survey of Drug Use and Health: 21.5%*
• Diagnosis by a “special specialist” was associated with higher

MAT uptake.

• Males diagnosed with OUD uptake MAT at higher rates than
• females. Why?

*Saloner B, Karthikeyan S. Changes in substance abuse treatment use among individuals with opioid use disorders in the United States, 2004-2013. JAMA. 2015 Oct 13;314(14):1515-7. Special specialist = Physician with specialty in psychiatry, addiction medicine, or pain medicine.

16

Summary

• More-recent years associated with lower MAT uptake. Possible

explanations:

• Increasing provider tendency to apply OUD diagnosis (so less-severe

patients get diagnosed) without concurrent increase in tendency to prescribe MAT

• Increasing tendency for OUD to be diagnosed by physicians not

authorized to dispense MAT

• “Depletion of the susceptibles” - Those most receptive to treatment

were already getting it early on

• Crowding of behavioral health services

17

Strengths and limitations of the study

Strengths Limitations

• Unmeasured confounding
• Population “at risk” of MAT is defined imprecisely
• Methadone dispensing at clinic is likely unobserved
• Opioids obtained without insurance involvement are unobserved
• Illicit opioid use is unobserved
• Large sample size from all regions of the U.S.
• Control for secular increase in OUD diagnosis when assessing
• ther barriers to MAT uptake
• Connect OUD diagnosis to MAT dispensing

18

Implications

Future efforts to increase MAT uptake may require increased involvement of, and patient access to, specialists in addiction medicine, pain management, and psychiatry. Greater efforts to train primary care physicians and

• ther specialists in implementation of MAT to treat OUD

may also reduce barriers to OUD treatment.

19

Thank you

Ezra Fishman, PhD

Contact: efishman@healthcore.com

Geetanjoli Banerjee, PhD John Barron, PharmD Gosia Sylwestrzak, MA

20

21

Appendix material

22

Full cohort inclusion/exclusion criteria

# Criterion Count

1 Starting Population: Any diagnosis code for opioid use disorder at any time 349,760 2 Index OUD Diagnosis in intake period 210,396 4 Exclude those with faulty Death Dates 210,305 5 Continuously Enrolled 12 Months Baseline 111,493 6 No Baseline OUD Diagnosis 102,544 7 Age LT 65 91,524 8 No Cancer 87,013 9 No Hospice 86,504 10 No Baseline MAT 73,406 11 Any Encounter in Baseline Period 70,438 12 Exclude those with faulty birth date 70,437

23

22% 14% 35% 22% 6%

Regional distribution of study cohort

MW NE S W Unknown

Percent of patients by region on index date

MW = Midwest, NE = Northeast, S = South, W = West.

24

Age pattern of OUD diagnosis in our cohort resembles age pattern in opioid overdose hospitalization rates

0.8 9.1 24.2 29.5 28 36.3 42.4 28.7 5 10 15 20 25 30 35 40 45 0-14 15-19 20-24 25-34 35-44 45-54 55-64 65+ Nonfatal overdose hospitalizations per 100,000 persons Age

Opioid overdose hospitalization rate, 2015, United States

Source: Centers for Disease Control and Prevention (CDC). 2018. Highlights from the 2018 Annual Surveillance Report

• f Drug-Related Risks and Outcomes - United States. Slide 31.

https://www.cdc.gov/drugoverdose/pdf/pubs/CDC_2018_Surveillance-Report_DataSummary_presentation.pdf Weighted national estimates from Healthcare Cost and Utilization Project Nationwide Inpatient Sample, 2015.

25

Patient characteristics by sex

Patient characteristics, by sex

Statistic FEMALE MALE P-Value Age < 0.0001 N 33,124 37,313 Mean (SD) 42.93 (14.03) 39.67 (14.41) Median 45 40 Min-Max 1-64 1-64 Region < 0.0001 MIDWEST 7,189 (23.11%) 8,656 (24.79%) NORTHEAST 4,099 (13.18%) 5,970 (17.10%) SOUTH 11,900 (38.25%) 12,550 (35.95%) WEST 7,920 (25.46%) 7,737 (22.16%) ECI < 0.0001 N 33,124 37,313 Mean (SD) 2.53 (2.36) 1.92 (2.12) Median 2 1 Min-Max 0-19 0-19 Index Year < 0.0001 2013 4,042 (12.20%) 5,375 (14.41%) 2014 5,091 (15.37%) 6,302 (16.89%) 2015 6,568 (19.83%) 7,609 (20.39%) 2016 8,308 (25.08%) 8,643 (23.16%) 2017 7,907 (23.87%) 8,082 (21.66%) 2018 1,208 (3.65%) 1,302 (3.49%) Index OUD Diagnosis Moderate Severe < 0.0001 N (%) 29,767 (89.87%) 32,548 (87.23%) Any Benzos Baseline < 0.0001 N (%) 11,725 (35.40%) 8,526 (22.85%) Baseline Opioid Rx Fill < 0.0001 N (%) 24,185 (73.01%) 22,719 (60.89%) Special Specialist in Baseline Prd < 0.0001 N (%) 14,117 (42.62%) 12,188 (32.66%)

Females tended to be older and have somewhat higher ECI. A slightly larger share of females had a “moderate/severe” OUD diagnosis (as opposed to “mild”). A smaller share of females were in the Northeast (the most-treated region) and a larger share were in the South (the least-treated region). A slightly higher proportion of females than males got their OUD diagnosis in later years. At the same time, females were more likely to have had an opioid Rx fill in the baseline period, more likely to have had a Rx fill for benzodiazepine, and more likely to have seen a “special specialist” in the baseline period.

26

Days to MAT uptake, among those ever initiating MAT in follow-up period, by year of OUD diagnosis

Statistic 2013 2014 2015 2016 2017 2018 N 3,160 3,430 3,559 3,428 2,641 359 Min 1 1 1 1 1 1 Median 21 29 28 19 8 4 P75 216 270 222.5 148 46 16 P90 670.6 688 529 367 148 36 P99 1,543 1,266 948 649 352 78 Max 1,862 1,563 1,138 817 436 116

27

All covariates in Cox model of MAT uptake

Covariate Estimated coefficient Standard Error Hazard Ratio Lower CI Upper CI z Value Pr(>|z|) Age

• 0.03

0.97 0.97 0.97

• 47.51< 0.0001

Sex: MALE 0.27 0.02 1.31 1.27 1.36 16.18< 0.0001 ECI

• 0.04

0.96 0.95 0.97

• 9.85< 0.0001

Region: NORTHEAST 0.14 0.03 1.15 1.1 1.21 5.6< 0.0001 Region: SOUTH

• 0.09

0.02 0.92 0.88 0.96

• 4.06< 0.0001

Region: WEST 0.02 1 0.96 1.05 0.12 0.9057 Special Specialist on Index Date 0.23 0.02 1.25 1.21 1.3 12.7< 0.0001 Any Benzos Baseline 0.25 0.02 1.28 1.24 1.33 13.31< 0.0001 Baseline Opioid Rx Fill 0.14 0.02 1.15 1.11 1.2 7.71< 0.0001 Index OUD Diagnosis Moderate Severe 0.43 0.03 1.54 1.46 1.62 15.93< 0.0001 Index Year: 2014

• 0.09

0.03 0.91 0.87 0.96

• 3.45

0.0006 Index Year: 2015

• 0.19

0.03 0.83 0.79 0.87

• 7.25< 0.0001

Index Year: 2016

• 0.29

0.03 0.75 0.71 0.79

• 11.18< 0.0001

Index Year: 2017

• 0.27

0.03 0.76 0.72 0.81

• 9.7< 0.0001

Index Year: 2018

• 0.18

0.06 0.84 0.75 0.94

• 3.14

0.0017

28

OUD severity, from DSM V (American Psychiatric Assoc., 2013)

ICD-10-CM Diagnosis code ICD-9 approximate equivalent Severity Number of symptoms F11.10 305.5x Mild (“abuse”) 2-3 F11.20 304.x Moderate (“dependence”) 4-5 F11.20 304.x Severe (“dependence”) 6+

Symptom list can be found here: https://www.asam.org/docs/default-source/education-docs/dsm-5-dx-oud-8-28- 2017.pdf?sfvrsn=70540c2_2

29

Appendix: abbreviations

Dx = diagnosis ECI = Elixhauser comorbidity index ICD = International Classification of Diseases MAT = medication assisted treatment OUD = opioid use disorder Rx = Prescription MW = Midwest, NE = Northeast, S = South, W = West