Data and Health Equity in Quality
Improvement
Renu Tipirneni, MD, MSc, FACP, Program Director
Selena Tran, MS, Health Data Analyst
July 18, 2025
Agenda
1. The Why
2. Shared Language
3. The Role of Data in Health Equity
4. Examples from Prior Literature
5. Group Engagement
Key Takeaways
➢ Health outcomes are heterogeneous
➢ Focusing on delivering high quality care for all patients means
understanding which groups may be experiencing different
outcomes
➢ Using race/ethnicity data is one way to understand how people
experience care
➢ Best place to start is where you are now—build out from there
The Why
Legacy of CQIs in Quality Improvement
Easy Wins
2026?
Heterogeneity in Outcomes around
Apparent Upward Trajectory
Shared Language
The world is changing rapidly…
…which means we must evolve
Health Equity / Health Care Equity /
Equity:
Definition / Synonyms
Synonyms to consider
• Ensuring access to high-quality care for all
patients
• Ensuring the opportunity for all to achieve a
better/high quality of life
• Achieving whole health* for all people
• Achieving the highest level of health for all
people
• Providing the opportunity for all people to
have/achieve optimal health
• Partnering together to ensure/help all people
to have optimal health and health care
outcomes
• Providing opportunities for better health and
health care choices
The fair distribution of health determinants, outcomes, and resources within and between
segments of the populations, regardless of social standing
*By whole health we mean physical, behavioral, spiritual, and socioeconomic well-being as defined by individuals, families,
and communities
The Role of Data in
Health Equity
“Failure to disaggregate race or ethnicity
data... can mask critical between- and
within-group differences that policies and
programs should address.”
Braveman, P., Arkin, E., Proctor, D., Kauh, T., & Holm, N. (2021). Systemic and
Structural Racism: Definitions, examples, health damages, and approaches to
dismantling. Health Affairs 41, NO. 2 (171-178)
Why do data
aggregation
choices
matter?
15%
17%
19%
21%
23%
25%
27%
2000 2005 2010 2015 2020
Adverse outcomes
(n = 1,150)
Why do data
aggregation
choices
matter?
15%
20%
25%
30%
35%
40%
45%
50%
2000 2005 2010 2015 2020
Adverse outcomes by subgroup
Group A (n=1000) Group B (n=150)
Race/Ethnicity as
Sociopolitical
Categories
The categories we collect
and report on are often
used as proxies for other
more difficult-to-measure
concepts
Race /
Ethnicity
Social
Drivers of
Health
Ethnic and
Cultural
Practices
Immigration
Status
Skin Color
and
Pigmentation
Social and
Stress-
Related
Biomarkers
Known
Ancestry
Data Collection
and Cleaning
Gather the
available race/
ethnicity data.
Clean and prep
data.
Information and
Insights
Stratify
performance
and health
outcomes by
race/ethnicity to
identify and
measure
disparities.
Analysis into
Action
Use insights to
develop QI
strategies and
interventions.
Intervention
Evaluation and
Adjustment
Share data and
information to
systematically
review and
evaluate
progress.
Adjust
intervention(s)
as needed.
Best Practices for Presenting Data in Visualizations
• Titles should correspond to the measure(s)
o "Performance by race/ethnicity, per hospital admission records"
o Alternatively, offer a definition of the measure as a footnote
• Include a footnote of the data source and the method used to collect the data
• Display the n under each bar label
• Display the x-axis categories/groups alphabetically
o "Prefer not to answer" and "Unknown" can be last in the order
• If there is missing data, include a footnote of how much missing data there is
(e.g., 30% missing)
• If groups are too small to report out, include this information as a footnote, but not
in the figure itself
1
Sample for Native Hawaiian or Pacific Islander was too small to report.
2
Of all patients, 10% do not have performance data.
3
Data was collected from 2015-2020.
75 75
80
75
78
75 75
65
0
10
20
30
40
50
60
70
80
90
American Indian
or Alaska Native
Asian Black or African
American
Hispanic or
Latino
Middle Eastern
or North African
White Prefer not to
answer
Unknown
Score
Performance by Race/Ethnicity, as determined by Patient Attributes File
n = 45 n = 45 n = 195 n = 90 n = 80 n = 1102 n = 30 n = 40
Avoiding Data Pitfalls
Do This
Compare to overall population or goal value
Disaggregate when possible; Address data limitations
Use non-
mutually exclusive race/ethnicity categories
Provide a rationale for inclusion of race/ethnicity as a
descriptor in any analysis
Avoid
Using White race as the default category
Collapsing small groups into “other” category
Using multiracial as an analytic category
Using race, ethnicity, and geographic origin as
proxies for genetic ancestry
Data Best Practices
Examples from Prior
Literature
MPOG study finds Black vs. White patient race was
associated with less antiemetic administration
• Previous studies found relationship between
socioeconomic status and antiemetic administration
• Model also adjusted for patient sex, hospital-level factors,
year, patient history (diabetes, motion sickness, smoking,
etc.), and specifics about the care team and anesthesia
technique
Standardized pathways may help reduce disparities
• ERAS Protocols:
o Standardized, multimodal perioperative
pathways
o Span the continuum of surgery (pre-, intra-,
post-operative)
o Include processes addressing patient
education, multimodal analgesia, early
mobility
o Multidisciplinary input and implementation
o Driven by best evidence, creates a culture of
pathway adherence
Group Engagement
Has your site been included in any departmental or
hospital initiatives related to health equity?
• If yes, what opportunities have you come across?
What challenges?
• If no, are there any you would like to be included in?
Is there any support needed to make this happen?
Key Takeaways
➢ Health outcomes are heterogeneous
➢ Focusing on delivering high quality care for all patients means
understanding which groups may be experiencing different
outcomes
➢ Using race/ethnicity data is one way to understand how people
experience care
➢ Best place to start is where you are now—build out from there
References
Braveman, et. Al. Systemic And Structural Racism: Definitions, Examples, Health Damages, And Approaches To Dismantling. Health Aff
(Millwood). 2022. doi:10.1377/hlthaff.2021.01394.
National Academies of Sciences, Engineering, and Medicine. 2025. Rethinking Race and Ethnicity in Biomedical Research. Washington,
DC: The National Academies Press. https://doi.org/10.17226/27913
Feero, et al. Guidance on Use of Race, Ethnicity, and Geographic Origin as Proxies for Genetic Ancestry Groups in Biomedical Publications.
JAMA. 2024. doi:10.1001/jama.2024.3737
Quint JJ, Keawe‘aimoku Kaholokula J. Now That We Are Disaggregating Race and Ethnicity Data, We Need to Start Understanding What
They Mean. JAMA Netw Open. 2024. doi:10.1001/jamanetworkopen.2024.3674
Frey T. Updated Guidance on the Reporting of Race and Ethnicity in Medical and Science Journals. AMWA. 2023.
doi:10.55752/amwa.2023.195
White, et. Al. Antiemetic Administration and Its Association with Race: A Retrospective Cohort Study. Anesthesiology. 2023. doi:
10.1097/ALN.0000000000004549.
