2. Environmental Public Health Tracking: Some Records are Vital Suzanne K. Condon, Director Bureau of Environmental Health Massachusetts Department of Public Health NAPHSIS/NCHS Joint Meeting, Salt Lake City, Utah June 5, 2007

3. Outline I. Environmental health and disease concerns II. Collaboration, communication and resource sharing with vitals partners * 24A * Provision of resources to enhance data quality of vital records III. Environmental Public Health Tracking: What have we learned to date? * IAQ & asthma * Lupus and 21e sites * DBPs (TTHMS) & low birth weight IV. Linking childhood cancer & birth outcome data with municipal water supply data V. Summary

4. Environmental Health & Disease Concerns Report of the Pew Commission: I. Environmental Health & Disease Concerns Report of the Pew Commission:  Source: Pew Environmental Health Commission 2000 Public Perception of Environmental Contribution to Health

5. Source: MDPH (extrapolated from Community Assessment Program Telephone Tracking System) Annual Calls Taken Regarding Perceived Environment and Disease Clusters

6. II.Collaboration, communication and resource sharing with vitals partners  24A M.G.L. c. 111, §24A authorizes the Commissioner of MDPH to approve “scientific studies and research which have for their purpose the reduction of morbidity and mortality within the commonwealth.” Receiving §24A approval protects any data or other information collected for purposes of the research as confidential.

7. II.Collaboration, communication and resource sharing with vitals partners  Provision of resources to enhance data quality of vital records Geo-coding of birth and death records

8. III. Environmental Public Health Tracking: What have we accomplished to date?

9. Density of Tier-Classified 21e Sites with Lupus-Suspected Chemicals and Neighborhoods with the Highest Rates of Lupus

10. Asthma Prevalence in Massachusetts Schools with Moisture/Mold Problems 1 2003/2004 - 2005/2006 School Years Prevalence 2 Asthma PrevalenceLow 0% - 8% Moderate 8.1% - 12% High 12.1% - 22% Total #%#%#% Schools with moisture/mold problems 1 2051.33274.41990.571 Schools without moisture/mold problems 1948.71125.629.532 Total 3 39100.043100.021100.0103 Chi square 10.8 p value < 0.01 1 Schools with moisture/mold problems are those where at least one classroom or library had water damaged ceiling tiles or carpet or visible mold growth on ceiling or carpet. 2 Prevalence of asthma in a school represents the prevalence during the school year during which the IAQ was assessed. 3 Excludes 5 schools that did not report prevalence data.

11.  Several epidemiological studies have shown increased risk of impaired fetal growth associated with exposures to disinfection byproducts (e.g., Bove et al. 1995, 2002 ; Infante-Rivard 2004; Savitz et al. 2005)  Third trimester of pregnancy: time during which fetal growth and birth weight may be most sensitive to environmental exposure opportunities (Kline et al. 1989) Scientific Purpose in Linking Birth Outcomes and DBPs in Drinking Water

12. Variables Available for Linkage  Environmental Data Water Quality Data  Health Data Low Birth Weight Data

13. Linkage Variables - Environmental Data  Drinking water quality data from municipal water supplies are housed in Massachusetts Department of Environmental Protection Water Quality System database  Water quality data from approximately 1988- present are computerized, with most complete data beginning in 1993 for public water supply (PWS) systems  This linkage project focused on total trihalomethane (TTHM) data

14. Linkage Variables - Health  Low birth weight data obtained from Massachusetts Registry of Vital Records and Statistics, housed in the Massachusetts Department of Public Health  Collect information on nearly 80,000 annual births in Massachusetts  Computerized since 1969 and available in Access database

15. Health Data Low birth weight data  Low birth weight defined as infant weight of less than 2,500 grams at birth  Key variables readily available include date of birth, maternal address at birth, birth weight, gender, some maternal risk factor information (e.g., smoking status)

16. Linkage Goals Assess low birth weight data in relation to nearest TTHM sampling location in community for possible relationships (i.e., greater occurrence of low birth weight babies in areas with higher TTHM levels)  Assign each live birth to nearest TTHM sampling location based on maternal address at birth  Assign TTHM level corresponding to the last trimester of pregnancy for each live birth  Calculate total low birth weight odds ratios according to TTHM exposure categories (≤ 40 ppb, >40-80 ppb, >80 ppb) across all communities

18. Results Linkage Analyses for Greater Boston Area  Communities served by the Massachusetts Water Resources Authority (MWRA), improves exposure assessment by relying on weekly, rather than quarterly, monitoring data (1999-2001). In addition, the water treatment regimens employed by the treatment system studies minimize the geographic variability of TTHMs  In addition to third trimester risk estimates typically derived in such studies, first, second and full pregnancy estimates were also calculated

19. Term Low Birth Weight & Exposure (all races N=780) Exposure1 st Trimester2 nd Trimester3 rd Trimester < 401.00 40 to <500.87 (0.72, 1.05)1.04 (0.84, 1.29)1.05 (0.79, 1.39) 50 to <600.90 (0.73, 1.11)1.12 (0.92, 1.49)0.84 (0.63, 1.11) 60 to 70+0.92 (0.73, 1.15)1.17 (0.92, 1.49)0.89 (0.66, 1.19) 70+0.87 (0.67, 1.13)1.31 (1.06, 1.66)0.96 (0.70, 1.32) Using logistic regression, a statistically significant increased risk for TLBW was found among those with high (70+ ug/L) TTHM exposure.

20.  Using logistic regression and adjusting for potential confounding factors, a statistically significant risk for term low birth weight was found among those with high (70+ ppb) TTHM exposure  To account for the potential impact of geographic variation in TTHM, a crude sensitivity analysis was conducted  Using existing exposure values assigned to each mother, the values were weighted on the basis of TTHM data from a single sampling site closest to the mother’s residence  This analysis raised the possibility that some unaccounted for geographic variation in TTHM data could lead to different exposure results and thus different overall results Results, cont.

21. IV. Linking Childhood Cancer & Birth Outcome Data with Municipal Water Supply Data  Overall Goal To assess opportunities for exposure to VOCs during fetal development and the risk of childhood cancer

22. Linkage  Cancer registry records and Registry of Vital Statistics records are linked to identify the maternal residence at time of birth for childhood cancer cases

23. Linkage Files  Registry of Vital Records Child’s name* Child’s DOB* Child’s gender Maternal residence at birth  MA Cancer Registry Child’s name* Child’s DOB* Cancer diagnosis Date of Diagnosis *Linkage Variables

24. Linkage Steps  Records in Cancer Registry and Vitals Registry each have a separate ID number  To maintain privacy of confidential cancer data, personal identifiers are not share with researcher  Using name and DOB, MDPH Privacy Officer matches Cancer Registry and Vitals Registry files, and assigns a unique Match ID number to matched records  Using the registry ID number and Match ID number, an analytic file is created that contains cancer type, DOB, and maternal residence (city/town, zip code) but not name  The analytic file is used by the researcher to assign exposure values and determine the risk of cancer while protecting privacy

25. V. Summary  Public remains concern about health impacts and possible environmental exposures  Use of vital records is often critical in responding to concerns  Linkage of existing data sets alone cannot answer complex environmental health questions  Data sharing agreements are important, but more important is the collaboration with data stewards and when possible the sharing of resources