Guiding Community Planning and Evaluation Efforts in Tobacco Control Using Data on Smoking During Pregnancy Vanessa Newburn Patrick Remington Paul Peppard.

Guiding Community Planning and Evaluation Efforts in Tobacco Control

Using Data on Smoking During Pregnancy

Vanessa Newburn

Patrick Remington

Paul Peppard

University of Wisconsin Population Health Program, Madison, WI

Objectives• Describe a method:

– Model progress in using local data– Set objectives which account for local variation

• Illustrate the method:– Birth certificate data from Wisconsin

• Recommendation:– Set objectives be set as a percent less than

predicted prevalence

Background

• Community-based programs are critical

• Few methods are available to guide local planning and evaluation efforts which:– Account for local variation– Can be systematically applied across

communities

• US Birth Certificates are a useful source of local data

Methods• Analysis Using Local Data:

– Calculate prevalence of smoking by year

– Model historical trends in prevalence

– Predict prevalence in a future year

– Set objective as a percent less than predicted

• Data (for illustration): – Wisconsin Standard Birth Certificates,

1990-2000

How to Model Local Trends

• Model:– Linear regression of log-prevalence on year

• Decreasing rates of change over time• Avoids predicting negative future prevalence

– Use slope to calculate Relative Annual Percent Change in Prevalence (100*(em-1))• Constant percent change from year to year • Describes progress over time

Model for the State--1

0

5

10

15

20

25

1990 1995 2000 2005

Percent smokers

3.3% per year relative decline

Model for the State--2

0

5

10

15

20

25

1990 1995 2000 2005

Percent smokers

3.3% per year relative decline

Model for the State--3

0

5

10

15

20

25

1990 1995 2000 2005

Percent smokers

Current State Objective

20% less than predicted

Variation Across Counties

CountyAverage No. Births/Year

Prevalence, 2000 (%)

Relative Change (%/Year)

95% CIPredicted

Prevalence, 2005 (%)

STATE (ST)

69025 16.3 -3.3 [-3.7, -2.9] 13.8

Variation Across Counties

CountyAverage No. Births/Year

Prevalence, 2000 (%)

Relative Change (%/Year)

95% CIPredicted

Prevalence, 2005 (%)

STATE (ST)

69025 16.3 -3.3 [-3.7, -2.9] 13.8

Kenosha (KE)

2089 17.7 -5.3 [-6.4, -4.1] 13.5

Waukesha (WK)

4127 9.4 -5.1 [-5.7, -4.4] 7.3

Outagamie (OU)

2186 14.1 -0.6 [-1.7, 0.6] 13.7

Shawano (SH)

469 22.8 0.5 [-2.5, 3.5] 23.3

Model for a County

0

5

10

15

20

25

30

35

1990 1995 2000 2005

Kenosha County

Percent Smokers

5.3% per year relative decline

Model for a County

0

5

10

15

20

25

30

35

1990 1995 2000 2005

Kenosha County

0

5

10

15

20

25

30

35

1990 1995 2000 2005

Waukesha County

Percent SmokersPercent Smokers

5.3% per year relative decline

5.1% per year relative decline

County Objectives

0

5

10

15

20

25

30

35

1990 1995 2000 2005

Kenosha County

0

5

10

15

20

25

30

35

1990 1995 2000 2005

Waukesha County

Percent SmokersPercent Smokers

20% less than predicted

20% less than predicted

20% less than observed

20% less than observed

Model for a County

0

5

10

15

20

25

30

35

1990 1995 2000 2005

Percent Smokers

Shawano County

0

5

10

15

20

25

30

35

1990 1995 2000 2005

Outagamie County

Percent Smokers

0.5% per year relative increase

0.6% per year relative decline

County Objectives

0

5

10

15

20

25

30

35

1990 1995 2000 2005

Percent Smokers

Shawano County

0

5

10

15

20

25

30

35

1990 1995 2000 2005

Outagamie County

Percent Smokers

20% less than predicted20% less than observed

20% less than observed20% less than predicted

Variation in Observed Prevalence Across Counties (n=72) in WI

02468101214161820

0 <3 6 9 12 15 18 21 24 27 30 33 36 54

WK KE SH

ST

OU

County prevalence of smoking during pregnancy, 2000 (%)

No. of Counties

Distribution of Relative Annual Percent Change in Prevalence

012345678910111213

ST

KE

WK

OU SH

No. of Counties

Relative Annual Percent Change in Prevalence

Limitations of Birth Certificates

• Collects data on smoking in most but not all states

• Data on women whose pregnancies result in live birth

• Underestimates prevalence (self-report bias)

• Link with overall smoking prevalence still being made

Application to Evaluation

• Within Communities:– Reflect on past progress– Set challenging objectives– Evaluate the impact of new efforts

• Across Communities:– Comparisons can be made if data is

consistent across jurisdictions

Summary of Results

• Substantial local variation exists:– Prevalence– Rates of change in prevalence– Local objectives should reflect variation

• Current methods for setting objectives are inadequate at the local level:– In WI, ¼ of counties & the State are

expected to meet 20% less than observed rate (2000)

Recommendations

• Use local data to model local trends – Data: birth certificates– Analysis: linear regression models of trends

in prevalence – Predict future prevalence

• Set objectives as a percent less than predicted

Acknowledgments

• Wisconsin Department of Health and Family Services– Kate Kvale, Randy Glysch

• Monitoring and Evaluation Program– Ann Christiansen, David Ahrens

• Wisconsin Tobacco Control Board• Funding: Agency for Healthcare Research

and Quality grant 5 T32 HS00083

Trends in Per Capita Cigarette Sales, Wisconsin and Arkansas, 1950-2000

30

50

70

90

110

130

150

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Year

Packs/y

ear

AR

WI

Trends in Per Capita Cigarette Sales, Wisconsin, 1950-2000

-30

-20

-10

0

10

20

30

40

50

60

Year

Packs/y

ear

CA

WI

AR

US

Trailer slide

• 24 slides x 1 set = 24 total slides

• Erich Mussak (265-9931)

• 1 set out of 1 in order

• Monitoring and Evaluation Program (WTCB)