Informing Medicaid Policy With Cancerrelated Health Services Research

1
Informing Medicaid Policy With Cancer-related
Health Services Research

• Siran M. Koroukian, Ph.D.
• Department of Epidemiology and Biostatistics
• Case Western Reserve University

2
Background

• Disparities in cancer-related outcomes by
  Medicaid status have been documented Medicaid
  beneficiaries are more likely than their
  non-Medicaid counterparts to
• be diagnosed with advanced stages of cancer
• to receive disparate cancer treatment and
  follow-up care
• gt to experience poor prognosis

3
Conceptual Framework

• Patient Sociodemographics
• Age, Race/Ethnicity, Sex,
• Insurance Status
• Medicaid Status

Cancer-related outcomes
Cancer Stage at Diagnosis
Access to and use of cancer screening services
Disparate cancer treatment and follow-up care
Access to and use of services for cancer
treatment and follow-up care
4
Insurance status, Medicaid status, and
cancer-related outcomes
Early Stage Diagnosis
Adequate access to and use of health services
Favorable outcomes/ good prognosis
Insured
Receipt of adequate care
Insurance Status
Uninsured/ Underinsured
Unfavorable Outcome/ Poor Prognosis
Advanced-stage disease
Inadequate access to and use of health services
HIGH OUT-OF-POCKET EXPENDITURES /
POVERTY/ RESOURCE DEPLETION
Receipt of disparate care/ Resource depletion
?
?
?
?
PARTICIPATION IN MEDICAID
5
Policy Questions

• Is Medicaid status associated with poor cancer
  (disease) related outcomes?
• gt Evaluate the effectiveness of the Medicaid
  program in cancer (disease) prevention and
  control
• CHALLENGING HYPOTHESIS PARTICIPATION IN MEDICAID
  IS ASSOCIATED WITH IMPROVED CANCER-RELATED
  OUTCOMES

6
Cancer as a CASE STUDY to examine policy-relevant
questions

• Difficulty to extract relevant disease
  information for other clinical entities to
  conduct policy analysis
• Availability of data from cancer registry ?
  information on when cancer was diagnosed, and the
  stage at which it was diagnosed (disease
  prevention)
• Cancer screening services
• Availability of well-established treatment
  protocols for some of the most common cancers ?
  comparisons between treatment received and
  guidelines, using claims data (disease control)
• Cancer treatment and follow-up care
• Quality of care
• Disease burden

7
Developing the linked Medicaid and Ohio Cancer
Incidence Surveillance System (OCISS)

• Linked database to mirror the SEER Medicare files
  at the Federal level, enabling the development of
  longitudinal records at the patient level to
  study patterns of enrollment in Medicaid and use
  of health services.
• Patient unique identifier in Medicaid to link
  enrollment and claims data across different time
  spans and service types.
• Linkage algorithm using patient identifiers
• Patient first and last name
• Date of birth
• Social security number
• Project approved by the Institutional Review
  Board at the Ohio Department of Health and by the
  Ohio Department of Job and Family Services

8
Description of the OCISS

• OCISS Mandatory reporting of all incident cases
  of cancer (except insitu cervical, squamous cell
  and basal cell carcinoma), since January, 1992
• Relevant data elements include
• Patient demographics
• Patient residence at the time of diagnosis
• Type of cancer
• Date of cancer at diagnosis
• Cancer stage
• Surgical treatment

9
Medicaid files

• Enrollment data ?
• Date of enrollment in Medicaid
• Length of participation in Medicaid prior to
  cancer diagnosis
• Claims data ?
• Health care utilization
• Screening
• Treatment (surgical radiation therapy
  chemotherapy)
• Follow-up care

10
Study 1 Assessing the effectiveness of Medicaid
in breast and cervical cancer prevention

• Analysis of cancer stage at diagnosis by timing
  of enrollment in Medicaid.
• Given that Medicaid is a safety net program,
  does it matter that we look at the timing of
  enrollment in Medicaid in relation to cancer
  diagnosis?

Koroukian SM. Assessing the effectiveness of
Medicaid in breast and cervical cancer
prevention. Journal of Public Health Management
and Practice, 2003 9(4) 306-314.
11
Figure 1 Proportion of women with advanced-stage
breast and cervical cancer at the time of
diagnosis, by Medicaid status
Diagnosed with Distant Metastases and 95
Confidence Interval
Medicaid Status
12
Figure 2 Proportion of women with advanced-stage
breast and cervical cancer at the time of
diagnosis, by Medicaid status, and by timing of
enrollment in the Medicaid program in relation to
cancer diagnosis
Diagnosed with Distant Metastases and 95
Confidence Interval
Medicaid Post-Diagnosis
Medicaid, Pre-Diagnosis
Medicaid Peri-Diagnosis
Non-Medicaid
Medicaid Status and Timing of Enrollment in
Medicaid
Peri-Diagnosis Women enrolled in Medicaid in
the 2 months prior to, upon, or in the 2 months
following cancer diagnosis Post-Diagnosis
Women enrolled in Medicaid 3 months after cancer
diagnosis
13
Study conclusions and implications

• Women enrolled in Medicaid shortly before, at, or
  after cancer diagnosis are significantly more
  likely to present with advanced-stages of the
  disease.
• Implications
• Methodological importance to account for timing
  of enrollment in Medicaid when identifying
  patients by Medicaid status
• Policy Medicaid as a safety net program.
  Could/should Medicaid reach out to the uninsured
  and the underinsured?
• Breast and Cervical Cancer Early Detection
  Program Who are the individuals presenting to
  Medicaid with advanced stages of cancer? Gaining
  better understanding of high risk populations and
  develop more effective targeting strategies for
  cancer screening.

14
Study 2 Does length of enrollment in Medicaid
matter that people receive cancer screening
services?

• Participation in the Medicaid program for the
  short term, and/or on a on/off basis does not
  benefit the patient
• Participation in Medicaid for the longer term may
  be associated with
• Continuity of care
• Increased likelihood to use services that are in
  the realm of preventive/screening/routine/follow-u
  p care

Koroukian SM. Length of Enrollment in Medicaid
Predicts the Use of Screening Mammography Among
Ohio Medicaid Beneficiaries. Accepted for
Publication, J Clin Epidemiol.
15
Mammography Screening in the Ohio Medicaid
Population by Length of Enrollment in Medicaid
Women 40-64 years of age, with no participation
in Managed Care programs, Medicare or spenddown,
and no stay in nursing homes after enrollment in
Medicaid
16
Women with At Least One Screening Mammography in
the 8-year Period, 1992-1999, by length of
enrollment in Medicaid
Women 40-64 years of age, with no participation
in Managed Care programs, Medicare or spenddown,
and no stay in nursing homes
17
Frequency of Screening Mammography in the 8-year
Period, 1992-1999, by length of enrollment in
Medicaid
Women 40-64 years of age, with no participation
in Managed Care programs, Medicare or spenddown,
and no stay in nursing homes after enrollment in
Medicaid
18
Average Number of Mammography Exams per Year by
Length of Enrollment in Medicaid
Women 40-64 years of age, with no participation
in Managed Care programs, Medicare or spenddown,
and no stay in nursing homes
19
Conclusions and study implications

• Increased length of enrollment in Medicaid is
  associated with greater likelihood to undergo
  screening
• Additional analysis needed to determine whether
  increased use of screening services reflects
  continuity of care
• Implications
• Methodological Importance to account for length
  of participation in Medicaid in studying use of
  screening/preventive services
• Policy Promote/facilitate sustained enrollment
  in Medicaid in order to enhance continuity of
  care

20
Cancer-related studies to inform Medicaid on
other methodological issues

• Ability of claims to identify incident cases of
  breast cancer (Koroukian SM et al. HSR Journal
  2003 38(3) 947-960).

21
Discussion

• Cancer as a case study. The findings are likely
  to also apply to other clinical entities.
• If participation of the underinsured and
  uninsured in the Medicaid program is associated
  with improved outcomes, perhaps consider
  proactive recruitment of individuals with
  potentially poor patterns of access to care into
  the Medicaid program.

22
Future Studies

• Gain a better understanding of the uninsured and
  underinsured populations in this case,
  individuals joining the Medicaid program upon
  being diagnosed with catastrophic illness.
• Study funded by an American Cancer Society grant
  underway to examine the characteristics of this
  population in association with community
  attributes, such as poverty and education.
• Gain a better understanding on the effectiveness
  of Medicaid in cancer prevention and control. A
  new study funded by the NCI (K07 CA096705) to
  examine differences in cancer-related outcomes in
  low-income Medicare beneficiaries is
  participation associated with improved outcomes?
• Cost burden to the Medicaid program
• Quality of care
• Access urban vs. rural
• Availability of health care resources (e.g.,
  radiation oncologists in association with breast
  conserving surgery hospital type and practice
  patterns in cancer treatment and follow-up care)

23
Acknowledgments

• The National Cancer Institute (F32 CA84621)
• American Cancer Society (IRG - 91-022-09)
• Collaborators
• Gregory S. Cooper, M.D.
• Alfred A. Rimm, Ph.D.