Data and Analytic Approaches for Studies of Health and Health Care Disparities Alan M. Zaslavsky John Z. Ayanian Harvard Medical School - PowerPoint PPT Presentation

Loading...

PPT – Data and Analytic Approaches for Studies of Health and Health Care Disparities Alan M. Zaslavsky John Z. Ayanian Harvard Medical School PowerPoint presentation | free to download - id: 6cd10a-MmZmN



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Data and Analytic Approaches for Studies of Health and Health Care Disparities Alan M. Zaslavsky John Z. Ayanian Harvard Medical School

Description:

Data and Analytic Approaches for Studies of Health and Health Care Disparities Alan M. Zaslavsky John Z. Ayanian Harvard Medical School – PowerPoint PPT presentation

Number of Views:66
Avg rating:3.0/5.0
Slides: 28
Provided by: AlanZ150
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Data and Analytic Approaches for Studies of Health and Health Care Disparities Alan M. Zaslavsky John Z. Ayanian Harvard Medical School


1
Data and Analytic Approaches for Studies of
Health and Health Care Disparities Alan M.
Zaslavsky John Z. Ayanian Harvard Medical School

2
Motivation
  • Much evidence of disparities
  • Health
  • Health care
  • Distinct issues and process
  • One of many sources of health disparities
  • What data are needed to detect disparities?
  • What data to understand and correct processes
    leading to disparities?
  • Conceptual approach to disparities
  • Analytic approaches for disparities

3
NAS/IOM Studies
  • Envisioning the National Healthcare Quality
    Report (2001)
  • Unequal Treatment Confronting Racial and Ethnic
    Disparities in Healthcare (2003)
  • Guidance for the National Healthcare Disparities
    Report (2003)
  • NHQR, NHDR now issued by AHRQ

4
  • IOM Unequal Treatment Report
  • Documented disparities in health care
  • Framework for definition and analysis of
    disparities

Institute of Medicine, 2003
5
NAS/CNSTAT data report
  • Eliminating Health Disparities Measurement and
    Data Needs (2004)

6
NAS/CNSTAT data report
  • Focus on collecting personal characteristics
    relevant to disparities research
  • Race/ethnicity
  • Socioeconomic position
  • Acculturation

7
Meaning of race/ethnicity
  • Socially-constructed groupings
  • Ethnicity common culture, origin, history
  • Race defined by putative genetic relationship,
    broad areas of origin
  • Many possible levels of detail
  • E.g. Hispanic vs. Salvadorean ethnicity
  • Asian race?
  • Inside vs. outside self-identification
  • Growing non-European immigration since 1965

8
Measuring race/ethnicity
  • Self-report as gold standard
  • OMB 1997 categories
  • 5 racial categories white, African-American,
    Asian-American, Native America/Alaskan Native,
    Hawaiian/Pacific Islander
  • Check all that apply format (2 multiracial
    in Census)
  • Hispanic/Latino ethnicity as separate question

9
Changes in categories
  • Before 1997 select one race
  • Standard allows for more detailed reporting
  • Must be collapsible to basic categories
  • More specific desired for state/local programs

10
Medicare race data
  • Primarily based on self-identification at Social
    Security enrollment
  • Little detail for older cohorts
  • See Arday et al. (HCFR 2000)

11
Comparisons to self report
  • White, African-American fairly accurate
  • White EDB sensitivity99, PV 93
  • Black EDB sensitivity95, PV 89
  • Less so for other groups (smaller, newer)
  • Hispanic sensitivity32, PV 93
  • Asian sensitivity42, PV 81

12
Collecting race/ethnicity at point of service
  • Self report versus staff report
  • Level of detail
  • Tailor to local populations
  • Avoiding duplication of effort
  • Sensitivity
  • Research and experience suggests acceptable if
    properly motivated

13
Socioeconomic Position (SEP)
  • Critical to disparities research
  • Mediator of disparities Race ? Lower SEP ?
    Poorer health care
  • Control variable to distinguish mechanisms
  • Source of disparities in itself
  • Complex interactions
  • E.g. SEP gradients for some outcomes are
    different across racial/ethnic groups

14
Dimensions of SEP
  • Current resources
  • Current income
  • Wealth/Assets (especially for elderly)
  • Permanent income ability to gain income
  • Education is an important measurable component
  • Occupation prestige, stress (UK research)
  • Life-course experience of deprivation/plenty

15
Collection of SEP data
  • Education
  • Relatively nonsensitive, simple to ask
  • Income
  • Highly sensitive in surveys, may be complex
  • Occupation
  • Less sensitive, complicated to code
  • Assets
  • Scales ask about key assets home, car

16
Acculturation/language
  • Complex concept characterizing meeting of
    cultures especially among immigrants
  • Ability to use health care systems
  • Discrimination based on foreignness
  • Protective and detrimental effects of
    culturally-specific practices
  • Changing expectations and needs
  • English language proficiency a key component
  • Barriers to communication and recognition

17
Acculturation measures
  • Place of birth, age at immigration
  • Generations since immigration
  • Language
  • Proficiency
  • Preference English or other language
  • Might be useful to providers/plans for
    communication, targeted outreach
  • Cultural identity scales

18
General issues
  • Broad range of data systems have different
    strengths and weaknesses
  • Coverage, sample size
  • Less for surveys
  • Level of detail and control over data collection
  • Less for administrative data
  • Many levels of detail possible
  • Different for research, federal and state programs

19
Linkages
  • Linkages across systems can bring together
    characteristics and outcomes
  • Linkages based on names or other keys
  • Geocoded census linkages contextual variables or
    approximate individual characteristics
  • Due regard to confidentiality concerns
  • Separate research and administrative uses

20
(Analytically) defining healthcare disparities
  • IOM Definition (Unequal Treatment)
  • Break race/ethnic differences into three parts
  • Due to health status-related variables
  • NOT part of disparity
  • Due to socioeconomic variables
  • Part of disparity
  • Remaining race/ethnic effects
  • Part of disparity

21
3 components
  • Health status-related variables
  • Age, sex, conditions, preferences, congenital
    susceptibilities
  • Socioeconomic variables
  • Income, education, employment/insurance
  • Remaining race/ethnic effects
  • Discrimination, statistical discrimination,
    poor communication

22
Operationalizing the IOM Definition
Health Status
Race
Utilization
SES
23
SEP Disparities
  • Analytic approach suggested by IOM Unequal
    Treatment report
  • Include SEP variables in models
  • Disparity Differences mediated through SEP
    Residual race effect
  • Differs from race coefficient in regression model

24
Respect for Preferences A Value or an Excuse?
  • Preferences may reflect
  • Personal and cultural values
  • Effects of past and present discrimination,
    patientss awareness of limited resources
    access, etc.
  • Ideal of informed preferences
  • Only attainable with adequate information,
    communication, access to care

25
Geographic Variation Alternative Treatments
  • Allow to mediate (like SES)
  • Geographical differences may be the result of
    historical patterns of oppression and
    discrimination.
  • Allows us to make comparisons across areas
  • Adjust (like health status)
  • Consider geography to be immutable if we are
    looking for improvements within an area
  • Consider geography as a preference an individual
    makes a decision to live in a given area

26
Multilevel analysis
  • Distinguish effects arise at various levels of
    healthcare system
  • Geography
  • Health plans
  • Providers (hospitals, doctors, clinics)
  • Patients
  • Effects of service patterns versus differential
    quality within units

27
Discussion questions (1)
  • Preferences
  • How would you distinguish between informed
    preferences and effects of past/current
    inequities?
  • What factors might contribute to apparent
    noncompliance?
  • Discrimination
  • Is it always conscious?
  • How would you prove discrimination?

28
Discussion questions (2)
  • Responsibility
  • Who is responsible for closing disparities
    entire system, or units that serve the most
    members of underserved groups?
  • Should we penalize or bolster underperforming
    providers who serve many minority patients?
  • How much responsibility does the health care
    system have for remedying the effects of social
    inequalities?
About PowerShow.com