Broward County EMA Medical Outcomes Pilot

1
Broward County EMA Medical Outcomes Pilot

• RWCA 2006 Grantee Conference Presenters
• William E. Green, Title I Grantee
• Michele Rosiere, Director Special Projects/
  Program Support
• Terri Sudden, Planning Council Coordinator
• Mark Young, Systems Administrator

2
Purpose of the EMA Medical Outcomes Project

• To develop a simplified and cost efficient
  companion database incorporating data from
  multiple sources to inform Planning Council
  decision-making including
• Comprehensive Plan
• Needs Assessment/Unmet Need Activities
• Priority-setting and Resource Allocation
• Evaluation Activities (including cost, outcomes,
  effectiveness, health disparities, and access)

3
Policy and funding are increasingly determined by
demonstrated outcomes...

• More emphasis is being placed on examining the
  relationship between expenditures and outcomes by
  the US Office of Management and Budget, the US
  Department of Health and Human Services (HHS),
  and Congress
• The ability of CARE Act programs to document
  outcomes depends on data gathering mechanisms and
  a carefully developed evaluation strategy

Source 2002 SPNS Report to CARE Act Grantees on
Thirteen SPNS Initiatives
4
Medical Outcomes
HRSA strongly recommends that grantees use
patient level health outcomes and corresponding
clinical indicators
Source Title I MAI Program Reporting
Instructions FY 2004
5
EMA MIS Evolution

• Current MIS System PROVIDE, Groupware
  Technologies Inc.
• FY 2004 2/3 of providers invoiced using MIS
• FY 2005 Provider contracts required invoicing in
  MIS
• FY 2006 Provider contracts required CD4 and viral
  load reporting in MIS
• Companion database fully operational in 2006
• Integrates longitudinal data including
• 2003-2005 invoices
• 2003-2005 CD4s and viral loads

6
Medical Outcomes Database System Design
7
Software Licensing/Vendor Support

• Uses Applications Service Provider (ASP) approach
  software
• A medical outcomes web-based database was
  designed to house data uploaded from a variety of
  software applications
• Including MS Excel, MS Access, and information
  downloaded from the EMA-wide patient and
  invoicing database system

8
System Design

• More than 400,000 records were imported from a
  SQL server database
• The majority of records included the EMA
  generated patient unique identifier (UI)
• UIs were generated for records that did not
  contain this information from the download
• Data integrity subroutines (including relational
  edits) were developed to clean data during the
  import process

9
Stakeholder Collaborations

• The Assessment Committee pre-determined a set of
  evaluation questions to begin analyzing
  patient-level medical outcomes and indicators
• Queries were developed in response to these
  pre-set questions
• Reports were generated and presented to the
  Assessment Committee throughout this process to
  provide an interactive feedback loop (Grantee,
  PC, Programs Support and MIS)
• More than 50 report formats were generated during
  the initial feedback process

10
HIPAA

• The Administrative Simplification provisions of
  the Health Insurance Portability and
  Accountability Act of 1996 (HIPAA, Title II)
• Requires HHS to establish national standards for
  electronic health care transactions and national
  identifiers for providers, health plans, and
  employers
• Addresses the security and privacy of health data
• Adopts standards to improve the efficiency and
  effectiveness of the nation's health care system
  by encouraging the widespread use of electronic
  data interchange in health care
• The Medical Outcomes Pilot process is fully HIPAA
  compliant

11
System/Software Requirements

• System requirements for this project include at a
  minimum
• Intel Xeon 5050 2x2MB Cache, 3.00GHz, 667MHz FSB,
  MS Server 2000, 2GB 533MHz (4x512MB), Single
  Ranked DIMMs, and, 80GB, SAS, 3.5-inch, 15K RPM
  Hard Drives (minimum of two hard drives)
• Software requirements
• Applications Service Provider (ASP) Approach
  Software
• MS Excel, MS Access, SPSS, and Crystal Reports
  (or other data analysis and reporting software)

12
Data Migration

• Data Collected Using Unique Identifier
• Example John Doe JHDE010119652
• 1st 3rd letter of 1st name, 1st 3rd letter of
  last name, DOB, gender 1 male, 2female, or
  3transgender
• Billing Data
• MIS System Generated Invoices
• Excel Generated Invoices
• Medical Chart Review
• CD4 and Viral Loads
• Invoice and medical data merged in MS Access

13
Medical Chart Review

• Unduplicated list of patients with a medical
  invoice in FY 03 or FY 04 by provider was
  generated
• EMA subcontracted with chart abstraction company
• Chart reviewers abstracted data elements over a
  two month period
• Abstracters entered data into web-based interface
• Over 3,000 unduplicated charts were reviewed from
  five medical providers

14
Chart Review Data Elements
15
Chart Review Data Web Interface
16
Data Limitations

• There is frequent crossover, even in individual
  patients, between funding sources such as Title
  I, Medicaid, AIDS Insurance Continuation Program
  (AICP), etc.
• Lab values were collected ONLY for Title I
  medical patients
• Participation in non-Title I medical care was NOT
  collected
• Medications for patients accessing non-CARE Act
  funded pharmacy services such as ADAP, AICP,
  Medicaid, Medicare, or the VA were NOT collected

17
Underlying Questions

• What services and combinations of services do
  patients utilize?
• Does participation in support services impact
  access and retention in medical care?
• Does receipt of support services impact medical
  indicators?
• What is the severity level of patients in care?
• Is there variability in patient severity by
  demographic categories in the system as a whole?

18
Unduplicated Patients by Service Category
19
What additional Title I-funded services do case
management clients use?
20
What percentage of Title I outreach clients
enter Title I-funded medical care?
Clients with Outreach invoices in FY 2003 were
selected and tracked through FY 2005 to determine
the percentage of clients by race and ethnicity
that entered Ryan White medical care.
21
Medical Outcomes Linked to Support Services
22
Severity Stratification

• National Institute of Health (NIH)
• Severe lt200 CD4
• Borderline 201350 CD4
• Moderate gt350 CD4 Viral Load gt100,000
• In Control gt350 CD4 and Viral Load lt 100,000
• Severity stratification is generated by data
  analysis and reporting software without
  developing complex flat files for statistical
  analysis (flat file development should occur
  after all data is assessed for uniformity)
• Provides system-level health status
• Informs disparity analysis
• Explores support services impact on health status

23
Severity Stratification by Overall Medical and
Support Category
A snapshot of patient severity stratification
was generated by selecting all patients enrolled
in a specific category by quarter with a CD4 and
VL from a RWTI medical provider. Mental Health
patients had a lower percentage of In Control
and Moderate stratification levels (62) than
clients in Oral Health (79). Analysis of
Variance is needed to determine if there are
significant differences between patients enrolled
by service category.
24
Appearance of Demographic Parity Across Gender,
Race and Ethnicity
There does not appear to be disparities in
medical severity by demographic variables.
Statistical analysis is needed to determine if
there are significant differences.
25
Increased In Control and Decreased Borderline
and Severe Stratification Frequencies for
Patients in Mental Health
This is a preliminary data run selecting clients
enrolled in medical care in Q1 and then entered
mental health in Q2. The preliminary results
point to a potential positive impact on clinical
indicators for those enrolled in primary care
prior to initiation of mental health services.
26
Increased In Control and Decreased Moderate
and Borderline Stratification Frequencies for
Patients Utilizing Dental Care
This is a preliminary data run selecting clients
enrolled in medical care in Q1 and then entered
oral health in Q2. The preliminary results point
to a potential positive impact on clinical
indicators for those enrolled in primary care
prior to initiation of oral health services.
27
Severe Clients Were Tracked to Inform System
Analysis
This is a preliminary data run selecting all
clients in the severe (lt200 CD4) stratification
category in Q1 and tracking change in severity
over 4 quarters.
28
Initial Invoice Trending was Performed
Invoice data for medical clients were selected to
trend monthly medical expenditures.
29
Future Data Collection and Analysis

• Implement requirement to report source of
  non-Title I medical care into MIS
• Implement requirement to report HAART in MIS,
  regardless of funding streams
• Ensure uniformly defined data elements
• Determine how to access data from other funding
  sources
• ADAP, Medicaid, Medicare, and VA

30
Broward EMA Reports Available www.BRHPC.org

• Special Population Access, Retention, and
  Adherence
• Haitian Population (2002)
• Recently Incarcerated Population (2003)
• Homeless Population (2004)
• Men Who Have Sex with Men (2005)
• Cost Effectiveness and Impact Evaluations
• Case Management on Primary Medical (2002)
• Mental Health and Substance Abuse (2003)
• Pharmacy Services (2004)
• Oral Health Care Services (2005)
• Quality of Life in the Fort Lauderdale/ Broward
  EMA (2000)
• Medical Outcomes (2001)
• Insurance Premium Purchasing Feasibility (2004)

31
Questions?

• William E. Green, Title I Grantee
• WGreen_at_Broward.org
• Michele Rosiere, Director of Special Projects
• MRosiere_at_BRHPC.org