carrie tomasallo, phd, mph wisconsin division of public health wisconsin asthma program
DESCRIPTION
Estimating Wisconsin Asthma Prevalence Using Clinical Electronic Health Records and Public Health Data. Carrie Tomasallo, PhD, MPH Wisconsin Division of Public Health Wisconsin Asthma Program [email protected]. Background. - PowerPoint PPT PresentationTRANSCRIPT
Estimating Wisconsin Asthma Estimating Wisconsin Asthma Prevalence Using Clinical Prevalence Using Clinical Electronic Health Records and Electronic Health Records and Public Health DataPublic Health Data
Carrie Tomasallo, PhD, MPHCarrie Tomasallo, PhD, MPHWisconsin Division of Public Wisconsin Division of Public HealthHealthWisconsin Asthma ProgramWisconsin Asthma [email protected]@wisconsin.gov
BackgroundBackground
Asthma is a prevalent chronic disease, Asthma is a prevalent chronic disease, affecting over 500,000 children and affecting over 500,000 children and adults in Wisconsinadults in Wisconsin
Wisconsin Behavioral Risk Factor Wisconsin Behavioral Risk Factor Surveillance System (WI BRFSS) data Surveillance System (WI BRFSS) data provide annual statewide asthma provide annual statewide asthma prevalence estimates prevalence estimates – data not useful for estimating prevalence data not useful for estimating prevalence
at smaller geographic areasat smaller geographic areas
Alternative Surveillance Alternative Surveillance DataData
UW Electronic Health (EHR) data from UW UW Electronic Health (EHR) data from UW Department of Family Medicine (DFM) Department of Family Medicine (DFM) Clinics to identify a patient population with Clinics to identify a patient population with asthma at a census block levelasthma at a census block level
Geographic analyses and maps may lead Geographic analyses and maps may lead to the identification and surveillance of to the identification and surveillance of Wisconsin asthmatic patients at Wisconsin asthmatic patients at neighborhood levelneighborhood level
Project GoalsProject Goals Can EHR data improve our estimate of asthma prevalence over telephone survey data?Can EHR data improve our estimate of asthma prevalence over telephone survey data?
How do How do aasstthhmmaa pprreevvaalleennccee eessttiimmaatteess bbaasseedd
on DFM clinic on DFM clinic ddaattaa aanndd BBRRFFSSSS ccoommppaarree??
Identify areas and populations of asthma disparity in Wisconsin using DFM clinic dataIdentify areas and populations of asthma disparity in Wisconsin using DFM clinic data
RationaleRationale
Current surveillance systems cannot Current surveillance systems cannot provide local level data within Wisconsin, provide local level data within Wisconsin, where many policies and interventions where many policies and interventions ultimately are designed and implementedultimately are designed and implemented
Use of EHR and socio-demographic data Use of EHR and socio-demographic data may improve on this method by accurately may improve on this method by accurately highlighting neighborhoods with high highlighting neighborhoods with high asthma prevalence in Wisconsin asthma prevalence in Wisconsin
These data may allow targeted education These data may allow targeted education and healthcare interventionand healthcare intervention
Limitations of WI BRFSS Limitations of WI BRFSS Asthma Prevalence Asthma Prevalence EstimatesEstimates
Designed for prevalence estimates at Designed for prevalence estimates at the national and state level but not local the national and state level but not local levels in Wisconsinlevels in Wisconsin
Small samples at county-levelSmall samples at county-level Even smaller samples for child estimatesEven smaller samples for child estimates Data obtained by self-reportData obtained by self-report Low response rates (~50%) may Low response rates (~50%) may
indicate response biasindicate response bias
BRFSS Asthma Prevalence by Wisconsin County 2007-2009
Clinical and Public Clinical and Public HealthHealthData ExchangeData Exchange
IRB approved limited data set of over IRB approved limited data set of over 195,000 patients (18,000 asthmatics) 195,000 patients (18,000 asthmatics) seen in UW Department of Family seen in UW Department of Family Medicine clinics in 2007-2009Medicine clinics in 2007-2009
Community partnership among Community partnership among clinicians (pulmonologist, primary care), clinicians (pulmonologist, primary care), population health scientists (Applied population health scientists (Applied Population Laboratory), and the WI Population Laboratory), and the WI Division of Public Health (Epidemiology Division of Public Health (Epidemiology & Public Health Informatics)& Public Health Informatics)
UW Department of Family Medicine Patient Population Location 2007-
2009Geographic Density of 195,000 Patients
Current Asthma Current Asthma DefinitionDefinition
BRFSS – Have you ever been diagnosed BRFSS – Have you ever been diagnosed with asthma? Do you still have with asthma? Do you still have asthma?asthma?
Clinical Data –asthma diagnosis (ICD-9 Clinical Data –asthma diagnosis (ICD-9
code 493) in encounter diagnosis or code 493) in encounter diagnosis or problem diagnosis fieldsproblem diagnosis fields
Child Asthma Prevalence 2007-2009
*Relative Standard Error > 30% (unreliable estimate)
Child Asthma Adjusted Odds Ratios 2007-2009
BRFSS model adjusted for sex, age, race/ethnicity and household income (BMI, personal smoking status or ETS exposure not available for children in BRFSS)Clinic model adjusted for sex, age, race/ethnicity, smoking status, BMI and census block level household income
Adult Asthma Prevalence 2007-2009
*Relative Standard Error > 30% (unreliable estimate)
Adult Asthma Adjusted Odds Ratios 2007-2009
BRFSS model adjusted for sex, age, race/ethnicity and household income, BMI, smoking statusClinic model adjusted for sex, age, race/ethnicity, smoking status, BMI and census block level household income
Clinic Patients with Asthma by Census Block
Group
ConclusionsConclusions Between 2007-2009, EHR clinic data Between 2007-2009, EHR clinic data
identified 18,000 asthmatics, compared identified 18,000 asthmatics, compared to 1,850 asthmatics from WI BRFSSto 1,850 asthmatics from WI BRFSS
BRFSS and clinic prevalence estimates BRFSS and clinic prevalence estimates and ORand ORadjadj were comparable were comparable
Clinic data had greater statistical power Clinic data had greater statistical power to detect associations, especially in to detect associations, especially in pediatric populationpediatric population
GIS analyses of clinic data identified GIS analyses of clinic data identified asthma patients at the census block asthma patients at the census block groupgroup
Future DirectionsFuture Directions Understanding where asthma Understanding where asthma
prevalence is highest and what prevalence is highest and what characteristics predict high characteristics predict high prevalenceprevalence
Method can be applied to any chronic Method can be applied to any chronic disease and other EHR data sets in disease and other EHR data sets in Wisconsin or U.S.Wisconsin or U.S.
Potential to address disparities by Potential to address disparities by identifying high risk communities to identifying high risk communities to target innovative interventionstarget innovative interventions
Collaborative EffortCollaborative Effort Brian Arndt-UW DFMBrian Arndt-UW DFM Bill Buckingham-UW APLBill Buckingham-UW APL Tim Chang-UW BiostatsTim Chang-UW Biostats Dan Davenport-UW HealthDan Davenport-UW Health Kristin Gallager-UW Pop HealthKristin Gallager-UW Pop Health Theresa Guilbert (PI)-UW Peds Theresa Guilbert (PI)-UW Peds Larry Hanrahan-DPHLarry Hanrahan-DPH David Page-UW BiostatsDavid Page-UW Biostats Mary Beth Plane-UW DFMMary Beth Plane-UW DFM David Simmons-UW DFMDavid Simmons-UW DFM Aman Tandias-DPHAman Tandias-DPH Jon Temte-UW DFMJon Temte-UW DFM Kevin Thao-UW DFMKevin Thao-UW DFM Carrie Tomasallo-DPHCarrie Tomasallo-DPH