Comparative Effectiveness of Virginia Coordinated Care versus the Traditional Safety Net Delivery System

Slide presentation from the AHRQ 2010 conference.

On September 27, 2010, Wally R. Smith made this presentation at the 2010 Annual Conference. Select to access the PowerPoint® presentation (1.2 MB). Free PowerPoint® Viewer (Plugin Software Help).


Slide 1

Comparative Effectiveness of Virginia Coordinated Care versus the Traditional Safety Net Delivery System

Comparative Effectiveness of Virginia Coordinated Care versus the Traditional Safety Net Delivery System

Wally R. Smith, MD, Principal Investigator

On behalf of
Virginia Commonwealth University
Medical Center
AHRQ MD-10-012

Slide 2

Background

Background

  • Uninsured:
    • Utilization patterns similar to Medicaid, underinsured:
      • Poor.
      • Lower social support.
      • Less transportation, education, delayed gratification.
      • May use ED rather than PCP.
  • Health Care Reform:
    • Reduces the number of uninsured, underinsured:
      • Of the 46 million uninsured, estimated 32 million will soon be covered.
    • Expands Medicaid program to approximately 16 million newly insured.

Slide 3

Do the Newly Insured Poor Change Utilization Patterns?

Do the Newly Insured Poor Change Utilization Patterns?

  • Not in some studies of providing managed care insurance-like programs to uninsured.
  • Barriers and Weaknesses of previous programs:
    • Short duration.
    • Dose of managed care variable:
      • Little case management.
      • Poor PCP (geographic or time) availability.
      • Few barriers to ED access.
      • Comorbidity mix unfavorable.

Slide 4

Virginia Coordinated Care 2000-present

Virginia Coordinated Care 2000-present

  • Patient-Centered Medical Home Insurance-like program for uninsured:
    • Uses managed care principles
  • All patients qualify for the Indigent care program supported by federal DSH and State General funds.
  • Primary care provided by community PCP's funded by VCUHS profits from commercial plans.
  • FFS and Management fee paid to PCPs in urban communities surrounding VCU:
    • Catchment area within 30 miles of VCU.
  • Patients given card with PCP's name.
  • Case managers support, assist with questions.
  • Enrollment files managed by Medicaid HMO owned by VCU Health System.

Slide 5

Initial Evaluation of VCC

Initial Evaluation of VCC

  • Lower ED visit rates.
  • Patients saw PCP.
  • VCC off-loaded patients to community physicians.
  • Community physicians happy with management fees.
  • Case management dose small.
  • Short-term evaluation only.

Slide 6

VCC Preliminary Data

VCC Preliminary Data

January 1, 2001, to December 31, 2003)

VariableNo. (%) N=2389
Before EnrolmentAfter Enrolment
Any inpatient discharge420 (17.6)330 (13.8)
Any emergency department visit1765 (73.9)1024 (42.9)
Any primary care visit557 (23.3)788 (33.0)
Any specialty care visit1729 (72.4)1895 (79.3)
P <.001 for all comparisons. 

Slide 7

VCC-Like Programs

VCC-Like Programs

  • NAPH national survey of safety net facilities:
    • 46 such programs at 37 facilities.
    • Most common motivation was to improve health and increase access (33%).
    • Many intended to reduce unnecessary care, improve satisfaction, offload clogged ED's.
    • Often urban academic medical centers.
    • Various catchment areas, target groups, structure.
  • In 2010, Medicare plans evaluation of existing multi-payer PCMH pilots:
    • Multi-Payer Advanced Primary Care Demonstration.
  • Do these programs work? If so, how?

Slide 8

Specific Aim 1

Specific Aim 1

  • Compare the effectiveness of the VCC program, particularly when there is active engagement with a VCC primary care provider, to the traditional safety net delivery system for reducing the frequency of emergency department utilization, hospital utilization, and adverse health outcomes, among indigent uninsured persons who used VCUHS from January 2003-December 2009, with attention to understanding the conditions and the types of patients for which the program was most effective.
  • 2003
  • 2003 VCUMC users
  • 2004 VCUMC users
  • 2005 VCUMC users
  • 2006 VCUMC users
  • 2007 VCUMC users
  • 2008 VCUMC users
  • 2010

Slide 9

Hypotheses, Aim 1

Hypotheses, Aim 1

  • Outcomes = ED utilization rates, hospitalization rates, charges, rates of ICU use, intubation, mortality:
    1. New VCC enrollees year following < year prior to enrollment.
    2. New VCC enrollees year following < uninsured not in VCC.
    3. Inversely correlated to # of visits to PCP:
      1. Year following enrollment.
      2. Three years following enrollment.
      3. Entire enrollment period (among intermittently enrolled).
  • Control variables = comorbidity, age, gender, race.
  • Subanalyze for AHRQ priority conditions of cardiovascular disease, diabetes, pulmonary disease/asthma.
  • Economic evaluation (costs, costs of program).

Slide 10

Patient Selection Aim 1

Patient Selection Aim 1

  • Age 18 to 63 years.
  • 2003-2009 utilization or enrollment.
  • Live in VCC-eligible geographic areas during analysis period(s).
  • Uninsured for at least a part of the study period.

Slide 11

Comparisons, Aim 1

Comparisons, Aim 1

  • Comparative analyses of utilization and adverse outcomes among various cohorts.
  • Comparisons between PCMHs (group practices) within the VCC.

Slide 12

Analytic Methods, Aim 1

Analytic Methods, Aim 1

  • Assemble several 2-year claims data cohorts (year prior and year after enrollment opportunities) grouped by VCC exposure.
  • Repeated measures methods to assess changes over time pre-post VCC enrolment.
  • Analysis of covariance to compare charge differences between males and females, various racial/ethnic groups, age groups and comorbidity, by exposure to VCC:
    • Consider models with interactions between these covariates and enrolee status
  • Logistic or Poisson regression for ED visits, hospitalizations.

Slide 13

Specific Aim 2

Specific Aim 2

  • Identify the successful and replicable elements of primary care case management, care coordination, and other attributes of the patient-centered primary care model that make the VCC program effective, with attention to understanding how and why these elements are critical for different patient populations.

Slide 14

Hypotheses, Aim 2

Hypotheses, Aim 2

  1. VCC administration, providers, case-workers, and patients perceive particular elements of VCC implementation, especially engagement with primary care providers, as important to the effectiveness of the program, and to improving its potential to be replicated in other settings.
  2. Particular attitudes, program designs and other actions by VCC administration may be important to the effectiveness of the program, and to the potential for replication.
  3. Patients with the most positive opinions about VCC and its program elements are those engaged with a primary care provider as well as those reporting the most medical need.

Slide 15

Data Collection and Analytic Plan, Aim 2

Data Collection and Analytic Plan, Aim 2

  • Qualitative analyses:
    • Environmental scan.
    • Key informant interviews.
    • Focus groups:
      • Informed by quantitative analyses
  • Survey draft:
    • For patients.
    • Based on Aim 1 results and Aim 2 key informant and focus groups results.
    • Draft for use in future surveys of program effectiveness.

Slide 16

Subject Inclusion Criteria, Aim 2

Subject Inclusion Criteria, Aim 2

  • VCC administration and leaders.
  • Key VCC providers.
  • VCC case workers.
  • Patients:
    • VCC enrollees.
    • VCC non-enrollees.

Slide 17

Personnel

Personnel

  • Aim 1:
    • Wally R. Smith, MD
      • Boling, Bohannon, Garland, Carcaise-Edinboro, Retchin, Bazzoli (economic eval), McClish.
  • Aim 2:
    • Alton Hart, MD, MPH
    • The National Public Health and Hospital Institute (NPHHI)
      • Linda Cummings, Sari Siegel-Spieler
        • Marshall, Aurich, Bernardo, Linson
  • Dissemination:
    • NPHHI

Slide 18

Importance and Impact

Importance and Impact

  • Examines "whether a promising system design, strategy, and intervention have actually improved the quality and efficiency of care and have affected related features of organizational or delivery system performance, including cost, access to care, reduction of disparities, and population-based health."
  • "Compares a design, strategy, and intervention that is likely to promote sustainable improvement in these desirable outcomes and is likely to improve performance on one or more dimensions of care quality..."

Slide 19

Relationship to IOM dimensions, AHRQ Priorities

Relationship to IOM dimensions, AHRQ Priorities

  • Level of analysis:
    • Organization.
  • Tests a system redesign.
  • Answers the IOM Initial Priorities for Comparative Effectiveness Research call for research on health care.
  • AHRQ priority populations including minorities and women constitute the bulk of the patients analyzed.
    • Initial national priorities for comparative effectiveness research. Institute of Medicine Web site. Accessed August 31, 2009.
Current as of December 2010
Internet Citation: Comparative Effectiveness of Virginia Coordinated Care versus the Traditional Safety Net Delivery System. December 2010. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/news/events/conference/2010/smith-w/index.html