Jail Medicaid Policy AnalysesFinal Report to the

Pinellas Data Collaborative

A Collaborative effort by The Policy Services & Research Data Center at the Florida Mental

Health Institute, Paul Stiles, Diane Haynes, Joe Morrissey, Hank Steadman, Kathy Dalton, &

Alison Cuellar

Funded by The MacArthur Foundation Network on Mental Health Policy Research

November 4, 2003

Two of the Network’s Goals

To improve access to quality mental health services

To provide research and practice to guide evidence-based policy

Our Study Objectives

Examine barriers to accessing services experienced by seriously mentally ill (SMI) people in the criminal justice systemEnsure that policies generated to reduce barriers are evidence-based

Estimated SMI Under Correctional Supervision in a

Given Day

Prisons Jails Community

Supervision

NumberOf SMIPersons

93,000 44,000 320,000

Current Policy Issues ~ 1

Growing interest by Federal agencies (DOJ/CHMS) in diversion & reintegration of mentally ill persons released from prisons and jails

Many jurisdictions are looking to Medicaid as a way of financing community services for these people

Current Policy Issues ~ 2 Lot of confusion and inconsistent

policies at Federal/state/local levels about disenrolling jail detainees from Medicaid, whether it’s required and whether it actually happens

Although research shows that Medicaid recipients have much higher access and use of general health services than the uninsured, there are no studies of this issue for SMI persons released from jail

Policy Research Question

Does Medicaid enrollment improve access and receipt of community-based services for

SMI persons released from jail?

Three Hypotheses

Higher percentage receiving services

Shorter lag times to first service More days of service and higher

rates of service receipt

Medicaid enrolled vs. not-enrolled within 90 days of jail release will have:

Study Design & Findings

Study Sites

Pinellas County (Florida): Main focus of current MacPolicy project; joint effort with FMHI-USF

King County (Washington): Leveraged earlier MacPolicy/current NIMH study to address same questions

Selected because of the availability of multi-agency administrative data:

Table 1. Year 2000 County Profiles

Indicator Pinellas County King County

Total Population 921,482 1,737,034

% White 87.3 75.7

% Black 9.1 5.4

% Asian 2.1 10.8

% Hispanic/Latino 4.7 5.5

Md. Income $37,111 $53,157

Avg. Jail Census 2,504 2,953

Total Jail Bookings 44,395 60,992

Incarceration Rate 4,818 3,511

Pinellas Data Collaborative Permitted by statute: Chapter 163, Part VI,

F.S.: Collaborative Client Information Systems.

Purpose: To establish an information system between state, local and private agencies to address various mental health planning and/or substance abuse planning.

Objective: to examine issues relating to how mental illness and substance abuse impacts the county in areas such as medical care funding, emergency medical transportation and the criminal justice system.

Available Databases for Pinellas

• Criminal Justice• Florida Department

of Law Enforcement• Baker Act (Civil

Confinements)• EMS• Agency for Health

Care Administration• Department of

Health

• CMS (Medicare)• Integrated Data

Systems (DCF)• Social Services• Juvenile Welfare

Board• Department of

Juvenile Justice• Child Welfare

King County Databases• King County Jail• King County

outpatient mental health services

• Medicaid enrollments

• State hospital services

• Community general hospital psych inpatient (CHARS)

• State Dept HSS substance abuse services (TARGET)

Sample Identification

Pinellas County: linked Medicaid claims (MediPass) with SMI identifiers to jail incarceration file for 2-yr. study interval

King County: linked county MH file with SMI identifers, Medicaid enrollment file, and jail incarceration for 2-yr. study interval

Case-Control Design

Cases: Persons w SMI released from jail with Medicaid enrollment

Pinellas: Jul98-Dec00 King: Jan97-Dec98

Controls: Persons w SMI released from jail without Medicaid enrollment

Sampling strategy controls for Medicaid eligibility—everyone was on Medicaid at some point during study interval

Sample Size & Medicaid Status by County

Sample PinellasMed

PinellasNonMed

PinellasTotal

KingMed

KingNonMed

KingTotal

TotalArrests 2,215 663 2,878 3,122 2,118 5,240

Total Persons 1,023 373 1,210 1,652 1,189 2,209

Pinellas County King County

Indicator

Medicaid

Non-Medicaid

Total

Medicaid

Non-Medicaid

Total

N % N % N % N % N % N % Age 35.6 35.2 35.3 35.7 Gender Male 1254 56.6 408 61.5 1662 57.7 1,081 63.7 860 69.4 1,941 66.1 Female 961 43.4 255 45.8 1216 42.3 615 36.3 380 30.6 995 33.9 Race/ Ethnicity

White 1479 66.8 483 72.9 1962 68.2 1,091 61.4 836 63.4 1,927 62.2 Black 522 23.4 128 19.3 650 22.6 602 33.9 413 31.3 1,015 32.8 Hispanic 9 <1.0 14 1.2 24 <1.0 NA -- NA -- NA -- Asian 1 <1.0 2 <1.0 3 <1.0 40 2.2 37 2.8 77 2.5 Other 202 9.1 36 5.4 238 8.3 45 2.5 33 2.5 78 2.5 Diagnoses Schizo- phrenia (295)

1013

45.7

268

40.4

1281

44.6

700

21.5

431

18.2

1,131

20.1

Affective (296)

1139

51.4

363

54.8

1502

52.2

2,074

63.4

1,516

63.9

3,590

63.7

Para- noia (297)

0

0

1

<1.0

1

<1.0

32

1.0

25

1.1

57

<1.0

Other Psycho- sis(298)

60

2.7

31

4.7

91

3.2

457

14.0

401

16.9

858

15.2

Demographic and Diagnosis Profiles of Jail Releases with SMI, Pinellas County (1998-00) and King County

(1997-98)

Service Access/Use Measures1. Percent received

any community-based service within 90-days

2. Days to first service

3. Days of service received within 90-days

4. Rate of service use within 90-days—adjusted for #days in community

Services Data Base

Pinellas County: used Medicaid claims and State block grant administrative data system (IDSDW)

King County: Medicaid claims not available; used county mental health and state substance abuse administrative data

Overview of Findings1. Two of 3 hypotheses confirmed—Medicaid cases

with SMI have higher access and more timely services following jail release but, once there, they receive a similar intensity of services

2. Medicaid advantage is not uniform across counties: Access: Pinellas 5:1 vs. King 1.25:1

(but non-M have 5:1 greater access in King)Lag-time: Shorter by 10-13 days in King CountyIntensity: Same for Medicaid/non-Medicaid in each county, but King County is twice as intense as Pinellas County

Average Service Use Indicators for Jail Releases with SMI by Medicaid Status, Pinellas and King

Counties Pinellas County King County

Service use indicatorsb/

Medicaid

Non-

Medicaid

Difference

Medicaid

Non-

Medicaid

Difference Probability of any service contact

.53 (2215)

.09 (663)

.44*** .60 (3122)

.48 (2118)

.12***

Days to first service contact

21.6 (1315)

29.2 (234)

-7.6*** 12.21 (1886)

16.54 (995)

-4.33***

Number of service days:

For users For all

9.56

(1315)

5.78 (2215)

10.38 (234)

3.97 (663)

-.82*

1.81***

20.08 (1886)

12.09 (3122)

18.82 (995)

9.30

(2118)

1.26

2.79***

Rate of service use: For users For all

.15

(1315)

.09 (2215)

.18

(234)

06 (663)

-.03

.03

.30

(1886)

.17 (3122)

.29

(995)

.13 (2118)

.01

.04**

a/ Based on duplicated sample of jail releases, not distinct persons.b/ Each average is adjusted for age, gender, race, diagnosis, length of incarceration, whether violent offense, and prior substance abuse treatment. Numbers in parentheses are sample sizes. *** p<.001, ** p<.01, *p<.05

Implications1. Medicaid is important for SMI

persons leaving jail2. But insurance is not the total

answer in promoting continuity of care (40-47% on Medicaid never accessed services in 90-days post release)

3. There was a stronger safety net in King County during late 1990s for the uninsured

Implications-cont’d4. Economic conditions post-9/11 have

eroded this safety net in King County and new CMS Medicaid rules (Jun ’03), if implemented fully, will do away with any service safety net for the uninsured throughout the US

5. Bottom Line: Big obstacles remain for persons with SMI in jail re: access to high quality, evidence-based services—these obstacles will intensify in next few years

Some Further Analyses1. Lengths of

detention for SMI persons

2. Why few SMI persons are disenrolled from Medicaid?

3. Factors that predict service use post-release among Medicaid enrolled

4. Charges for SMI persons—what’s the mix of minor & serious offenses?

Next Steps

Examine the impact of Mental Health and/or Substance abuse Treatment in reducing jail

recidivism