Transcript Predictors of Buprenorphine Adoption in Methadone and non

Predictors of Buprenorphine
Adoption in Methadone and nonMethadone Treatment Settings
Lori J. Ducharme, Ph.D.
Hannah K. Knudsen, Ph.D.
Paul M. Roman, Ph.D.
University of Georgia
Organizational Resources &
Evidence-Based Practice
• Much recent research has examined the links between
organizational resources and the adoption of evidencebased practices, including pharmacotherapies, in
addiction treatment.
• Among the predictors of medication adoption are
measures of:
– program quality (e.g., accreditation),
– workforce professionalism (e.g., employing physicians, counselor
certification and education),
– philosophical orientation (e.g., 12 step vs. medical model
programs), and
– caseload characteristics (e.g., diagnostic groups).
• However, less attention has been paid to program
revenue streams as potential predictors.
Organizational Resources and
Medication Adoption
• Understanding program resources demands a more
nuanced consideration of revenue sources than simply
program ownership (i.e., “public” vs. “private”).
• The extent of program reliance on commercial insurance
has been positively associated with the adoption of
pharmacotherapies such as naltrexone and disulfiram.
• By contrast, programs that are more dependent on
Medicaid and other public revenue streams have been
less likely to adopt pharmacotherapies.
• In these analyses, we examine the relative contribution
of these revenue streams to models predicting program
adoption of buprenorphine.
Data Sources
• Data were collected in face-to-face interviews
with administrators of 991 community-based
treatment programs.
• Pooled data from 3 samples are used:
– Nationally representative sample of 401 private-sector
treatment units
– Nationally representative sample of 350 public-sector
treatment units
– 240 units affiliated with NIDA’s Clinical Trials Network
• Organizational data were collected in 2003-’04.
• Data on adoption of buprenorphine were
collected via brief telephone follow-ups at 6
months after onsite visit.
Descriptive Statistics for Study Sample
Treatment programs in sample
N=991
Adopted buprenorphine at 6-month follow-up
13.2%
Government operated
11.0%
For-profit
16.6%
Accredited
52.4%
Methadone unit
18.4%
Offers detox services
26.6%
Employs physician(s)
43.5%
% master’s level counselors (mean)
44.5%
Program size (in FTEs, mean)
33.9
% primary opiate clients (mean)
22.3%
% revenues from Medicaid (mean)
13.1%
% revenues from commercial insurance (mean)
15.8%
Bivariate Associations: Organizational
Characteristics and Buprenorphine Adoption
Adopters
(13.2% of 991)
Non-Adopters
(86.8% of 991)
Government operated*
2.6%
13.2%
For-profit
20.0%
16.5%
Accredited*
79.1%
50.0%
Methadone unit*
30.7%
17.1%
Offers detox services*
50.9%
22.1%
Employs physician(s)*
63.5%
41.3%
% master’s level counselors (mean)*
50.8%
43.9%
45.4
31.3
% primary opiate clients (mean)*
34.9%
20.4%
% revenues from Medicaid (mean)
11.3%
13.7%
% revenues from commercial insurance (mean)*
29.8%
13.6%
Program size (in FTEs, mean)*
* indicates p<.05
Multivariate logistic regression predicting
buprenorphine adoption (n=831)
Model 1
Odds ratios
Model 2
Odds ratios
Government owned
.248*
.253*
For-profit
.971
.858
Accredited
1.780*
1.549
.742
.987
Offers detox
2.875**
1.996*
% Primary opiate clients
1.015**
1.018**
Methadone unit
Employs physician(s)
1.507
% Master’s counselors
1.950*
1.598
FTEs (log transformed)
1.067
1.148
% Medicaid revenues
---
.984*
% commercial insurance revenues
---
1.016**
*p<.05, **p<.01
1.610
(p=.06)
Interpretation of Results
• Both models show the influence of a number of
organizational characteristics on the adoption of
buprenorphine.
• The addition of revenue variables in Model 2 attenuates
the significance of the effects for accreditation and the
percentage of Master’s level counselors.
• Each of the revenue variables significantly predicted the
likelihood of buprenorphine adoption in this sample.
– A standard deviation (20.5%) increase in the percentage of
revenues obtained from Medicaid decreases the odds of
buprenorphine adoption by 27.9% (27.9=100*[e(-.016)(20.5) -1]).
– A standard deviation (23.2%) increase in the percentage of
revenues obtained from commercial insurance increases the odds
of buprenorphine adoption by 44.9% (44.9=100*[e(.016)(23.2) -1]).
Results, cont’d
• The proportion of clients with primary opiate dependence
also remains a significant predictor in Model 2, net of the
effects of revenue sources.
– A standard deviation (26.8%) increase in the proportion of primary
opiate clients increases the odds of buprenorphine adoption by
57.7% (57.7 = 100*[e(.017)(26.8) -1]).
• Net of the effects of revenue sources and caseload,
buprenorphine adoption was twice as likely among facilities
offering detox services, and 75% less likely among
government-owned programs.
• Employing at least one physician has a positive effect that
approaches statistical significance (p=.062).
Implications for OTPs
• OTPs have been viewed as being at a “disadvantage” relative to other
modalities regarding the adoption of buprenorphine, due to the
restrictive regulatory requirements for dispensing bup/nx in OTP
settings.
• In the bivariate analyses, OTPs were more likely than other programs
to have adopted bup/nx. However, multivariate models found that, net
of other organizational variables and revenue sources, OTPs were
neither more nor less likely to have adopted buprenorphine.
• On the one hand, these findings suggest that OTPs are not
disadvantaged by key organizational or resource issues to a greater
extent than other modalities.
• On the other hand, the lack of resource differences suggests that
adoption of bup/nx in OTPs could be significantly greater if the
regulatory restrictions were lessened.
Future Directions
• In general, the inverse association between Medicaid
and buprenorphine adoption is due to the lack of
inclusion (through 2004) on Medicaid formularies.
– However, Medicaid may also be an indicator of the
socioeconomic status of a program’s caseload.
• Conversely, the positive association between commercial
insurance and buprenorphine adoption may indicate a
greater likelihood of reimbursement, or may be serving
as a proxy for socioeconomic status of program clients.
• In either case, programs with a relatively more affluent
client base are likely to have greater access to resources
that facilitate the adoption of pharmacotherapies and
other resource-intensive technologies.
• Further research is needed on the regulatory and policy
environments in which organizational decisions about
service offerings are enmeshed.
The authors gratefully acknowledge the support of
research funding from the National Institute on
Drug Abuse (R01DA13110 and R01DA14482).
For more information and related publications, visit
the National Treatment Center Study website at
www.uga.edu/ntcs.