Evaluation of Maternal Smoking Surveillance Systems in Massachusetts Lizzie Harvey, MPH CDC/CSTE Applied Epidemiology Fellow Massachusetts Department of Public Health June 14, 2011 1 Disclosure • No significant financial interest or other relationships with the manufacturer(s) of any commercial product(s) or provider(s) of any commercial services discussed in this presentation and with any commercial supports of the activity • Massachusetts Department of Public Health RaDAR and PRAMS clearance 2 Objective • To evaluate maternal smoking surveillance through the Massachusetts Pregnancy Risk Assessment Monitoring System (PRAMS) 3 Importance of Maternal Smoking Surveillance • Maternal smoking during pregnancy is associated with babies who are: – 1.4-3.0 times more likely to die of Sudden Infant Death Syndrome (SIDS) – at 30% higher odds of premature delivery – 2.3 times more likely to deliver term low birth weight infants 4 Maternal Smoking Relevance in US and Massachusetts • Healthy People 2020 Objective – Increase smoking cessation during pregnancy • CDC Winnable battle • National Performance Measure – Smoking in the last trimester • MA Priorities – Improve the health and well being of women in their childbearing years – Support reproductive and sexual health by improving access to education and services • Opportunity for Intervention in MA – Tobacco Cessation and Prevention Program 5 Existing Surveillance System: Massachusetts Birth Certificate (BC) 6 MA Birth Certificate • Maternal smoking questions since 1986 • Since 1996, two questions: 7 Maternal Smoking in Massachusetts: Birth Certificate Data % MA Women Reporting Smoking during Pregnancy on the Birth Certificate 15 13.1 12 9 7.5 6.9 6 3 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 MA PRAMS 8 initiation Massachusetts Pregnancy and Risk Assessment Monitoring System (MA PRAMS) as a Surveillance System 9 PRAMS Background Population-based data on maternal attitudes and behaviors, before, during, and shortly after pregnancy 10 MA PRAMS Background • Initiated in 2007 • 80 questions (54 Core, 16 Standard, 10 MA developed) • 2-6 months post-partum • Administered in English and Spanish only 11 MA PRAMS Screener 3 months Last 3 months before pregnancy of pregnancy Current use 12 Maternal Smoking Data Comparison Pre-Pregnancy 12 mo. Pregnancy Conception 3 mo. Post-partum Delivery 3.8 mo. BC PRAMS 13 Methods • Data Sources: – 2007-2008 MA PRAMS (N=2,997) – Linked to 2007-2008 MA BC data (N=2,997) • Analysis: – Frequencies, prevalence estimates, sensitivity, positive predictive value, kappa coefficients, chi square – SAS 9.2 and SUDAAN 10.0 • Reference: – CDC: Updated Guidelines for Evaluating Public Health Surveillance Systems (2001) 14 CDC: System Attributes • • • • • • • • • Simplicity Flexibility Data Quality Acceptability Sensitivity Positive Value Predictive Representativeness Timeliness Stability 15 Simplicity MA BC • Entire population (~75,000 births/year) • DPH Parent Worksheet with 2 smoking questions • 49 licensed birth hospitals in MA MA PRAMS • Population-based mixed-methodology survey (~1,500/year) • 4 questions on 80 question survey • Mail survey to stratified sample of birth population 16 Flexibility MA BC • Iterations of maternal smoking surveillance – 1986 – 1996 – 2011 • 2011 implementation of electronic 2003 standard birth certificate MA PRAMS • Maternal smoking questions Core • Compare to other states • Opportunity to select standard and statespecific questions • Other measures related to smoking knowledge and behaviors 17 Data Quality Maternal Smoking Surveillance Data Quality: Missing Fields, 2007-08 % Missing 3.00 2.70 2.60 2.10 2.10 Last 3 Months Now 2.00 1.00 0.25 0.24 0.30 0.30 Cig Use Prior Cig Use During Cig Use Prior Cig Use During 0.00 BC Pre Merge BC Post Merge N=156,734 N=2,997 Screener 3 Months Prior PRAMS N=2,997 18 Acceptability MA BC • Mandatory participation – General Law (Ch. 111, s.24B) • Stigma on maternal smoking MA PRAMS • Not required by law – Can refuse survey and refuse individual questions • Infant outcome at 4 months post-partum may determine how mother will respond 19 Sensitivity Analysis Smoking DURING Pregnancy MA BC MA PRAMS S NS S 113 79 192 NS 20 2714 2734 133 2793 2926 Sensitivity: 113/ (113 + 20) = 0.850 MA PRAMS captured 85% of all maternal smoking during pregnancy identified by the BC 20 Predictive Value Positive (PVP) Smoking DURING Pregnancy MA BC MA PRAMS S S NS NS 113 79 192 20 2714 2734 133 2793 2926 PVP: 113/ (113 + 79) = 0.589 58.9% of all maternal smoking during pregnancy identified by MA PRAMS were cases identified by the birth certificate 21 But wait…the BC is not a gold standard! True maternal smoking rate unknown Kappa Coefficients measure agreement between categorical items taking chance into account Value of K Strength of agreement < 0.20 Poor 0.21 - 0.40 Fair 0.41 - 0.60 Moderate 0.61 - 0.80 Good 0.81 - 1.00 Very good 22 Kappa Coefficient Analysis Maternal Smoking DURING Pregnancy N=2,997 20 BC 113 79 PRAMS Κ = 0.68 (95% CI = 0.62 – 0.74) Good Agreement 23 Kappa Coefficient Analysis MA BC Non Quitter Smoker Non smoker 2497 98 MA PRAMS # cigs # cigs NS dec. sa/inc. resumed 32 27 -- Κ = 0.53 95% CI: Quitter 32 70 14 5 0 0.49 – 0.57 Moderate # cigs dec. -- 10 58 34 0 # cigs sa/inc. -- -- 9 11 0 NS resumed 0 0 0 -- 0 --: 1-4 values suppressed Agreement 24 Representativeness % MA Women Smoking during Pregnancy 15 12 Combined, 10.2 PRAMS, 9.3 9 BC, 6.9 6 3 0 1996 1998 2000 2002 2004 2006 2008 2010 Year 25 Representativeness: Maternal Smoking by Data Source 25 18.8 17.6 % of Mothers 20 15 10 14 ↑34% 12.6 10.2 9.3 7.2 ↑42% BC PRAMS Combined 5 0 Smoked Before Smoked During Smoked PostPregnancy Pregnancy Pregnancy 26 Representativeness: Differences in Reporting Maternal Smoking DURING Pregnancy 20 BC 113 79 PRAMS Who is reporting on PRAMS but not on BC? 27 Demographic Characteristics of those who report maternal smoking during pregnancy on PRAMS (N=192) Characteristic % PRAMS only % PRAMS and BC Chi-square p-value White Race (vs. non white) 63.3 65.5 0.75 Black Race (vs. non black) 21.5 19.5 0.72 ≥HS grad 73.4 77.0 0.57 ≥College 32.9 40.7 0.27 ≥ 30 years 26.6 29.2 0.69 Hispanic Ethnicity 35.4 17.7 0.005 Spanish Language 13.9 1.8 0.002 Married 24.1 23.9 0.98 Had Pre-pregnancy insurance 49.4 46.9 0.74 WIC 77.9 69.6 0.21 LBW 7.6 14.2 0.16 28 Representativeness MA BC • Population data • Stigma of maternal smoking • Self-report • Recall bias MA PRAMS • Stratified sampling by race/ethnicity – 30% non response • English and Spanish only • Stigma of maternal smoking potentially decreased • Self-report • Recall bias 29 Timeliness MA BC MA PRAMS DPH Parent Worksheet completed PRAMS Survey completed 49 Registrars Data Entry Data Entry Limiting Step Limiting Registry of Vital Records and Statistics Birth File Closed MDPH Birth Report Data weighted by CDC Public Health Action! PRAMS Report Step 30 Public Health Action! Stability MA BC • Required by law • 2003 Revised Birth Certificate Implementation 2011 • Investment in Vital Information Partnership System (VIP) 2011 • MassCHIP MA PRAMS • 5 year funding approved 2011-2016 – Decreased funding than previous cycle • PRAMS website • PONDER (MA specific) and CPONDER 31 Conclusions: MA PRAMS System Attributes Simplicity: 4 questions Flexibility: Ability to add state specific questions Data Quality: Low % missing; survey data Acceptability: Overall stigma potentially decreased Sensitivity: Increased case Kappa statistic ascertainment Positive Predictive Value: Representativeness: Higher ascertainment of maternal smoking; Additional measure of post-partum smoking Timeliness: No real time data; 2 year lag Stability: 5 year competitive funding from CDC (2011-16) 32 Recommendations • Use PRAMS as a valuable data source in addition to BC to understand burden of maternal smoking • Use PRAMS to fill the data gap regarding maternal smoking in post-partum period • Use other PRAMS data to inform actionable interventions in maternal smoking Prenatal patient education Postpartum smoking environment 33 Next Steps • Add 2009 MA PRAMS data to analysis • Assess impact of new BC data on maternal smoking • Continue work with the Tobacco Cessation and Prevention Program – Identify women who are not reporting on the birth certificate but reporting in PRAMS – Reach out to prenatal providers for universal screening of maternal smoking behaviors and 34 referrals to cessation programs Acknowledgements Hafsatou Diop, MDPH Office of Data Translation Karin Downs, MDPH Bureau of Family Health and Nutrition Thomas Land, MDPH Tobacco Cessation and Prevention Program Emily Lu, MDPH PRAMS Coordinator Alice Mroszczyk, MDPH Privacy and Data Access Office Maria Vu, MDPH Registry of Vital Records and Statistics CDC/CSTE Applied Epidemiology Fellowship Program 35 References/Resources CDC Tobacco Use and Pregnancy: http://www.cdc.gov/reproductivehealth/TobaccoUsePregnancy/index.htm (Accessed 6/2011) CDC Winnable Battle: Tobacco: http://www.cdc.gov/WinnableBattles/Tobacco/index.html (Accessed 6/2011) CDC PRAMS: http://www.cdc.gov/prams/ (Accessed 6/2011) CPONDER: http://www.cdc.gov/prams/CPONDER.htm (Accessed 6/2011) MA PRAMS: http://www.mass.gov/dph/prams (Accessed 6/2011) MassCHIP: http://www.mass.gov/dph/masschip (Accessed 6/2011) MA General Laws regarding birth information collection: http://www.malegislature.gov/Laws/GeneralLaws/PartI/TitleXVI/Chapter111/Section24B (Accessed (6/2011) 2003 Revised Birth Certificate: http://www.cdc.gov/nchs/data/dvs/birth11-03final-ACC.pdf (Accessed 6/2011) CDC Updated Guidelines for Evaluating Public Health Surveillance Systems: http://www.cdc.gov/mmwr/preview/mmwrhtml/rr5013a1.htm (Accessed 6/2011) Kappa Statistics: Cohen, J. (1968). Weighted kappa: Nominal scale agreement provision for scaled disagreement or partial credit. Psychological Bulletin, 70(4), 213. Dietz, PM, et al. Infant Morbidity and Mortality Attributable to Prenatal Smoking in the U.S. Am J Prev Med 2010;39(1) 45-52. 36 Questions? Contact Information: Lizzie Harvey, MPH CDC/CSTE Applied Epidemiology Fellow Massachusetts Department of Public Health Lizzie.Harvey@state.ma.us (617) 624-5559 37 Extra Slides 38 Resources in MA • Massachusetts Smokers’ Helpline • QuitWorks 39 Additional PRAMS Questions on Maternal Smoking New 2009 Core: Which of the following statements best describes the rules about smoking inside your home now? Check one answer No one is allowed to smoke anywhere inside my home Smoking is allowed in some rooms or at some times Smoking is permitted anywhere inside my home 40 2003 Standard BC 41 Impact of 2003 BC Pre-Pregnancy 12 mo. Pregnancy Conception 3 mo. Post-partum Delivery 3.8 mo. BC PRAMS 42 Demographic Characteristics of those who report maternal smoking before pregnancy on PRAMS (N=372) Characteristic % PRAMS only % PRAMS and BC Chi-square p-value White Race (vs. non white) 58.9 65.4 0.20 Black Race (vs. non black) 22.8 15.9 0.09 ≥HS grad 79.1 81.3 0.60 ≥College 42.4 45.3 0.57 ≥ 30 years 26.6 29.2 0.69 Hispanic Ethnicity 29.8 17.8 0.007 Spanish Language 9.5 1.87 <0.001 Married 32.9 28.5 0.36 Had Pre-pregnancy insurance 59.4 54.9 0.38 WIC 63.2 63.4 0.98 LBW 8.9 9.4 0.87 43 Sensitivity Analysis Smoking BEFORE Pregnancy MA BC MA PRAMS S S NS Smoking DURING Pregnancy MA BC MA PRAMS NS S NS 214 158 372 S 113 79 192 40 2499 2539 NS 20 2714 2734 254 2657 2911 133 2793 2926 Sensitivity: 214 / (214 + 40) = 0.842 Sensitivity: 113/ (113 + 20) = 0.850 MA PRAMS captured 84.2% of all maternal smoking prior to pregnancy and 85% of all maternal smoking during pregnancy identified by the BC 44 Predictive Value Positive Smoking BEFORE Pregnancy MA BC MA PRAMS S S NS Smoking DURING Pregnancy NS S 214 158 372 40 2499 2539 254 2657 2911 PVP: 214 / (214 + 158) = 0.575 MA BC MA PRAMS S NS NS 113 79 192 20 2714 2734 133 2793 2926 PVP: 113/ (113 + 79) = 0.589 57.5% of all maternal smoking prior to pregnancy and 58.9% of all maternal smoking during pregnancy identified by MA PRAMS were cases identified by the birth 45 certificate Kappa Coefficient Analysis Maternal Smoking Maternal Smoking PRIOR to Pregnancy DURING Pregnancy N=2,997 N=2,997 40 BC 214 158 PRAMS Κ = 0.64 (95% CI = 0.60 – 0.69) Good Agreement 20 BC 113 79 PRAMS Κ = 0.68 (95% CI = 0.62 – 0.74) Good Agreement 46 Representativeness: Differences in Reporting Maternal Smoking Maternal Smoking PRIOR to Pregnancy DURING Pregnancy 40 BC 214 20 158 PRAMS 113 BC Who is reporting on PRAMS but not on BC? 79 PRAMS 47