Analysis of Access of Services for Those Dealing with Substance Abuse Diane Haynes, M'A' Ren Chen, M

1
Analysis of Access of Services for Those Dealing
with Substance AbuseDiane Haynes, M.A.Ren
Chen, M.S.Policy, Services and Research Data
CenterMental Health Law PolicyFlorida Mental
Health InstituteUniversity of South
Florida1/21/03
Pinellas Data Collaborative
2
Introduction

• A study looking at individuals with substance
  abuse problems and patterns of interactions with
  services and governmental agencies in Pinellas
  County

3
Topics of Discussion

• Recap criteria of identification
• Todays Presentation
• Cluster Analysis
• Four Groups Identified
• Findings

4
Criteria of Identification

• Medicaid, DSS, EMS
• Alcohol and/or Substance Abuse diagnosis
• CJIS Had at least two arrest on statutes dealing
  with DUI, DWI, Under Influence, Disorderly
  Intoxication, or Inhalation of harmful chemical
  substance.

5
Criteria of Identification cont..

• IDS
• Alcohol and/or Substance Abuse diagnosis
• Cost Center of Methadone or Substance Abuse Detox
• Program of Adult or Children Substance Abuse
• Service Code of Assess DUI/DWI or Day
  Care/Additions

6
What method to use?

• How to identify patterns across multiple systems?
• Why Cluster analysis?
• What key information to use?

7
Why Cluster analysis?

• It is a multivariate analysis technique to
  organize information about variables so that
  relatively homogenenous groups, or clusters,
  can be formed.

8
What key information to use?

• Variables that will identify patterns of system
  interactions and intensity of interactions
• A variable for each system and then number of
  times they interact with that particular system

9
Four Clusters Identified

• Cluster one
• 1,247 individuals
• Low interaction with EMS system
• Largest cluster
• The highest interaction w/ any system was
  Medicaid and IDS
• Worst Outcomes
• 908 individuals
• Highest interaction w/any system was with CJIS
• Low interaction w/service agencies

• Cluster Three
• 363 Individuals
• They are receiving services from
  IDS/Medicaid/Child Welfare systems
• High interaction with Baker Act system
• Best Outcomes
• 120 Individuals
• Smallest cluster group
• Low interaction with CJIS system
• High interaction /IDS,Medicaid, Child Welfare

10
Four Clusters Identified
11
Cluster Demographics
12
Gender
13
Race
14
Age
15
First System Interaction
16
Second System
17
Arrested in FY 2000-2001
18
Clusters Adults vs Children
19
Cluster OutcomesAdults vs Children
20
Conclusions

• Individuals mostly only hit one system, so there
  is significantly low interaction across multiple
  systems
• The cluster analysis identified four distinctly
  different groups (NOT using individual
  characteristics), but interestingly there are
  significant characteristic differences in the
  individuals between the cluster groups .

21
Conclusions cont.

• Cluster 1, which is the largest group
• Cluster 1 is mostly White adult males who hit
  CJIS system first and then only a small portion
  hit MED, IDS as their second system.

22
Conclusions cont.

• Cluster 2 had the highest proportion of
  individuals hitting the CJIS system
• Cluster 2 had the worst outcomes
• Individuals were more likely to be white adult
  males

23
Conclusions cont.

• Cluster 3 Individuals first and second system
  hits were spread out across all systems
• They were slightly older than the individuals in
  the other clusters, had almost equal proportion
  of females to males.

24
Conclusions cont.

• Cluster 4 was the smallest group and did not hit
  the CJIS system
• They were significantly younger than those in the
  other clusters, and tended to hit MED, IDS, or CW
  for their first system
• Their second system if they had one was either
  MED, IDS,or DSS
• They had better outcomes than the other clusters

25
Further Questions

• Is cluster four just a younger population and in
  the future will they fall into the other
  clusters?
• What is the cost per individual across each
  cluster?
• Do geographic patterns using Zip Code in GIS
• Why do most individuals only hit one system over
  a three year period, no matter which cluster they
  are in? How many, and which systems should an
  individual dealing with Substance Abuse hit?

26
Whats Next - Options

• Continue with these four groups and ask
  additional questions
• Use different methods to identify groups
• Chose a cluster and focus further analysis, case
  review, etc. on individuals