Transcript Access and Consumption of Substance Abuse

Access and
Consumption of
Substance Abuse
Services
Academy Health Washington D.C.
Robert Wood Johnson Foundation’s
New Connections Program
Juliette Roddy, PhD
University of Michigan Dearborn
June 9, 2008
Problem


Behavioral Health lags behind the
medical mainstream in measurement
of evidence based practice
Behavioral Health lags behind the
medical mainstream in application of
evidence based practice
Problem

Provision of Behavior Health Issues:
• Lack of agreement on best practices
• Limitations on treatment funding regardless of
whether the behavior is ceased
• Clinics that specialize in specific care

Methadone treatment, abstinence, 12 step
Particularly in the field of substance abuse,
practice has been guided more by preferences
and training of the providers than the research
results on effectiveness
(Miller, Aweben and Johnson 2005)
Problem


There is a recognizable gap between
diffusion of research, innovation, and
implementation to practice in substance
abuse treatment (Miller 2006)
There is enough research for experts to
rank evidence based treatments according
to strength and type of evidence (Miller,
Aweben and Johnson, 2005)
Broad Questions

The Robert Wood Johnson Foundation’s
Addiction Prevention and Treatment
team (renamed) identified tracking the
use of evidence-based treatment
approaches as one of its top priorities
1.
2.
3.
4.
5.
Brief screening for substance abuse in primary care
The use of pharmacology to treat addiction
Psychosocial interventions in specialty care settings
The variety and prevalence of multi-systemic
treatment
Follow up care
Specific Questions




How often are adult consumers given brief
screening interventions in the primary care
setting?
How often are adults offered pharmacological
solutions to treat addiction/abuse disorders?
Among adults engaged in outpatient treatment,
how many are offered the services of:
psychosocial interventions, wraparound
services, and aftercare services?
What was the consumers’ opinion on the
services that were offered? Were the services
effective?
Proposed Study


Answer the specific questions in terms of
frequency
Use two separate existing data sets (not
pooled)
• Fighting Back Data National Comparison Set
• National Survey of Alcohol, Drug and Mental
Health Problems

Investigate influences on consumption of
services through correlation, regression,
factor and cluster analysis
Data




National Survey of Alcohol, Drug and
Mental Health Problem (NSADMH)
Part of RWJF’s Health Tracking Initiative
Designed to track changes in American
health and healthcare
A broad national survey with questions on
demographics, drug use, use of
medications, and access and utilization of
behavioral healthcare services
Data






NSADMH, year 2000
N= 12,158
Mean age 49 (min 18, max=98)
Nationally representative
Phone Interview
Restricted Use
Data



National Evaluation of the Fighting Back Program
Specifically associated with RWJF’s community
based drug abuse prevention program and the
national comparison survey
Survey collected information on health and
mental health, knowledge and utilization of AOD
treatment services, attendance in drug education
courses or lectures in school. Background
variables include sex, race, household
composition, marital/cohabitation status,
education status and achievement, employment
status, occupation, religious preference, and
income.
Data






NFB, year 1999
N= 3,297
Mean age 31 (min 16, max=44)
Nationally representative
Phone Interview
Publicly available
Methods


Begin with descriptive statistics that
can be identified as direct measures
that answer the research questions
Examine correlations of a variety of
variables to identify interesting
relationships


Eight separate categories of insurance
provision were included in the FB Data
Many insurance categories were correlated
with types/categories of substance abuse
care
Methods

Regression analysis was used to
verify the correlation results
• Consumption of SA services examined
for significant components
• Insurances entered together and
separately (ie Medicare and Medigap)
• Betas from regressions would be
unreliable, however significance could
be determined
Methods


Cluster analysis examined for
unifying characteristics
Factor analysis scattered and difficult
to interpret
Results
Abbreviated Correlation Matrix rf
Currently
Insured
Covered
by emplyr
Medicare
Medigap
Medicaide
Military
Insurance
Other
public
plan
Insurance
bought on
own
Insurance
covers SA
TX for SA
in last 12
months
Hospitalized
For SA
care
Residential
Inpatient
care
Emergency
Room
Treatment
.024**
-.143**
-.065
.004
-.095*
-.040**
-.043**
.094**
.009
(phi)
Attended
SA self
help
group
-.115**
PCP
gave
Meds
for
SA
-.026**
-.092**
-.206**
-.037**
-.043**
-.055
-.009
.016
.066
-.014
-.028
.089**
.007
.065*
.011
.131**
.043
-.017
-.030**
.087**
.012
-.017
-.030**
.077**
.021*
.057**
.096*
.052
.138**
.043**
.018*
-.022*
-.023
-.037
-.045
-.028**
-.018*
.061**
.129
.018
.005
.024
.039**
*significant at the .05 level
**significant at the .01 level
-.036**
Results
Regression Results
Dependent Variable Treatment for MH/SA past 12 months
Variable
Constant
Gender
Employer Insurance
Medicare
Medigap
Medicaid
Military Insurance
Other public insurance
Other insurance plans
Other self purchased plan
PCP asked about SA
Beta
t-statistic
.056*
.006*
-.033*
-.027*
.091*
.015
.049*
-.005
.002
.150*
7.056
5.525
0.463
-2.624
-2.279
8.392
1.440
4.797
-0.468
.217
14.881
The insurance variables were entered together and separately to verify significance.
Betas may not be reliable.
Adjusted R2 = .05, F statistic 41.1
Results
Clusters (N=1311)
Covered by Employer
provided insurance
# of months of SA care
past 12 months
Attended Self help in
past 12 months
PCP referred to SA
specialist
PCP gave SA
medication
PCP spent >=5 mins on
SA counseling
Cluster 1 (n=37)
1
Cluster 2 (n=19)
0
3
11
1
0
0
1
0
1
0
1
Results
1. How often are adult consumers given brief screening interventions in the
primary care setting?
NSADMH data: N=12,158
Received SA treatment in past 12 months: 1413 (11.6%)
Had PCP visits past 12 months? n=9965 (82%)
PCP asked about substance abuse? n=2425 (20%)
PCP referred patient to SA specialist n=417 (3.4%)
FB data: N=3297
Received treatment or counseling in past 12 months? 42 (1.3%)
2. How often are adults offered pharmacological solutions to treat
addiction/abuse disorders?
NSADMH data: N=12,158
PCP gave SA medication: 1027 (8.4%)
3. Among adults engaged in outpatient treatment, how many are offered the
services of: psychosocial interventions, wraparound services, and aftercare
services?
FB
FB
FB
FB
FB
FB
data: N=167 (ever) received services (valid percents given)
data: n= 90, 45 (50%) Received a physical as part of SA care
data n=87, 50 (58%) Had a mental health evaluation as part of SA care
data n=90, 11 (12%) rec’d employment counseling as part of SA care
data n=90, 29 (32%) had relationship counseling as part of SA care
data n=9961, 417 (4.2%) were referred to a MH/SA specialist by PCP
4. What was the consumers’ opinion on the services that were offered? Were
the services effective?
NSADMH data: N=12,158
Respondent reports got less treatment than needed: 548 (4.5%)
Satisfaction with SA Care: 3656 (69.5%)
Dissatisfied with SA care : 171 (1.5%)
Discussion

Analysis suggests insurance influences
access and provision of treatment for
substance abuse disorders
• Significant in correlations, regressions and
clusters
• Hardly new information
• 2001 RWJF reported that 2/3 of the funding for
alcohol and drug treatment is from public
sources with Medicare and Medicaid paying for
a full 21 percent of treatment services
Discussion

Often assumed that providers of insurance
will not cover treatment that is ineffective
and that efficient treatments will be
covered by insurance
• Not necessarily the case
• Literature has expressed that absence of
insurance coverage for treatment does not
always indicate that a treatment is neither safe
nor effective
• Furthermore, the fact that treatment is
covered by insurance does not ensure that it is
effective (Steinberg and Luce, 2005)
Discussion

Evidence based treatment standards
and cost considerations will continue
to impact treatment for behavioral
health
• Does evidence based suggest integrity
of care?
• Evidence ranges from gold standards
(double blind clinical trials) to anecdotal
case reports
Discussion



Considerations other than the strength of
evidence may also be useful
Impact of the treatment, relevance to
subject population, and the consequences
of treatment could all be considered when
examining efficacy
It is also important to note absence of
evidence for efficacy does not equate to
ineffective treatment
Conclusions

Insurance influences both access to
and provision of care
• Behavioral healthcare standards are
changing
• Reimbursement of care is slowly being
tied to outcomes (Miller et al 2005)
• Insurers will require evidence (and
there will be a direct relationship
between strength of evidence and
allowance)
Future Research/Lessons Learned

Data is limited
• Without full investigation it becomes
very difficult to understand it the care
offered was evidence based or not
• (I even understand the APT team’s
questions better now)
Must explore probit or logit analysis for
the regresstion. These are linear.
Thanks!
Dr. Debra Joy Perez
Dr. Margarita Alegria
Catherine West
Dr. Allen Goodman