Transcript Diapositiva 1

Computing for a Cure:
Why Mammogram Rates in Utah are Lower
than the National Average
Noralynn Valletta
Mentored by Dr. Cecily Heiner
COSE Symposium, November 12th, 2012
The Problem
• Breast cancer is the leading cancer
killer of women in Utah. (Ahmedin
Jemal, 2010)
• Utah’s mammography rates are one
of the lowest in the nation. (Kaiser
Family Foundation, 2010)
Early Detection
Mammograms can reduce breast cancer mortality
rates by increasing early detection.
(Komen)
Why Aren’t Women Getting
Mammograms?
Possibilities:
• Financial implications
• Lack of formal education
• Proximity to health care location
Hypothesis:
Rural women are less likely to participate in mammogram
screenings due to lack of education.
Data
Start with survey data provided by the
Center for Disease Control (CDC’s)
Behavioral Risk Factor Surveillance System (BRFSS).
BRFSS is a long survey with questions about
• Drinking
• Age
• Household members
• Visits to doctors
• Dozens of questions
Analysis
• Goal: Find statistically significant
variables to aid our understanding of why
women may not utilize the early detection
that mammograms provide.
• Tool: SPSS software – statistical software
• Analyisis: Binary logistic regression
– Outcome variable based on HADMAM
• Whether or not they had a mammogram
• Binary variable- either they did or they did not!
Glossary
Statistically significant
variables are important
because they reflect a
pattern rather than mere
coincidence.
• SPSS (“Statistical Package
for the Social Science”)
software
• Data Mining: attempt to
discover patterns in large
data sets
Variable Mapping
Query Variable
BRFSS Name
Statistical Significance
Dental Plan
DENCLEAN
YES
Education
EDUCA
NO
Ownership of health plan
HLTHPLAN
YES
Religious beliefs
DRNKANY4
NO
County
CPCOUNTY
sometimes
UT County Map By
Mammogram Rate
Pink: abnormally low
mammogram rates
Blue: less than 10
individuals surveyed
Purple: Both!
Mobile Screening:
“The Mammogram Truck”
• Mammogram
screenings brought
to your area
• Available to local
businesses or
groups of 10+
women
More Information:
(801)-713-0600
Future Work
• Recently funded by the Utah Komen Affiliate
• Design and develop software with an adaptable survey
to pinpoint reasons behind low mammogram rates in UT
• Offer facts, statistics and real-life stories to motivate and
educate participants; follow up with a quiz one week later
Thank You!
References
• Ahmedin Jemal, R. S. (2010). Cancer Statistics, 2010. CA: A Cancer
Journal for Clinicians, 277-300.
• The Henry J. Kaiser Family Foundation. (2010). Mammogram Rate
for Women 40 years+. Retrieved September 18th, 2012, from
http://statehealthfacts.org/comparetable.jsp?ind=479&cat=10&sub=
113&yr=138&typ=2&sort=a
• http://ww5.komen.org/uploadedFiles/Content/GetInvolved/Events/M
ammo_Eng.pdf
• http://www.cdc.gov/nchs/data/hus/hus11.pdf#090