Surveillance Blind Spots

The Davenport Project · APHA 2026 Annual Meeting · San Antonio, Texas

Surveillance Blind Spots: Geographic Clustering of Congenital Cardiovascular Defects in the Illinois-Kentucky Fluorspar Belt

Anahied Rayzian, BS/MPH Student, Department of Public Health, Northern Illinois University
Dr. Tomoyuki Shibata, Ph.D., College of Health and Human Sciences, Northern Illinois University

Full poster accompaniment, to explain and expand on methodologies and results.

Overview

This study follows congenital heart defect (CHD) surveillance from the nation to one Illinois county and describes the gaps at each level.

National. Incomplete and inconsistent national data leave the true CHD burden uncertain [1, 2, 3, 7].
State. Illinois CHD rates cluster, and EPA-tracked exposures do not account for the pattern [8–16].
County. Pope County has the highest rate in Illinois, but small numbers and a state line keep it from being distinguished from chance [8, 12, 20–23].

Background

Congenital heart defects (CHDs) are the most common birth defect in the United States, affecting nearly 1% of births, or about 40,000 a year [2]. Birth defects tracking in the U.S. is run by individual state programs whose methods vary [3].

The Illinois-Kentucky Fluorspar Belt (geologically, the Illinois-Kentucky Fluorspar District) is a former fluorite mining district. In Illinois it lies in Pope and Hardin Counties; in Kentucky, in Crittenden, Livingston, and Caldwell Counties [4, 22]. Its ore also carries galena, barite, and sphalerite, which are lead, barium, and zinc minerals [4]. Fluorspar mine tailings at the Minerva Mine #1 in Hardin County contain barium, cadmium, and lead above comparison values; the 2007 health consultation concluded that current exposures there posed no apparent public health hazard [5]. Both Illinois counties lie in the Shawnee Hills karst region [6].

Methods

Data: NBDPN, 2014–2018 [1]; Illinois APORS county rates and births, 2016–2020 [8, 13]; EPA TRI, Superfund, and AirToxScreen [14–16]. Analysis: Moran's I, LISA, Kulldorff scan, and Spearman correlation [9–12]. Limits: county-level data, diagnoses rather than infants, Illinois data only, and Pope selected for its high rate.

National coverage

This study looks at the 50 states and the District of Columbia. The NBDPN Birth Defects Data and Directory (2014–2018) lists 37 programs [1]. Pennsylvania reported no heart defect data, the Department of Defense program covers a non-geographic population, and Puerto Rico is outside the 50-state scope, so 34 state programs were included (12,743,197 live births). For each program, the published prevalences of 16 cardiovascular defect categories were summed into CHD diagnoses per 10,000 live births. d-TGA was excluded because it is a subset of TGA. Published prevalences were used because some programs report some defects for fewer years. Coverage is program live births divided by 19,559,660 U.S. births for 2014–2018 [7].

Illinois clustering and exposures

County rates of major cardiovascular defects in children under 2 come from IDPH APORS, 2016–2020, Table 5 [8]. Global Moran's I used queen contiguity with row-standardized weights on Census 2020 county boundaries (R package spdep) [9, 10]; rook contiguity was run as a check. Local Moran's I (LISA) identified local clusters at p < 0.05 [11]. A Kulldorff Poisson spatial scan (R package smerc, 999 Monte Carlo replications, seed 20261103, windows up to 10% of Illinois births) used IDPH county birth counts for 2016–2020 as the denominator [12, 13]. Settings were fixed before the scan was run.

Spearman correlations compared county CHD rates (n = 102) with seven EPA measures: TRI facilities, total releases, water releases, and metal releases for 2016–2020 [14]; Superfund NPL sites, all statuses [15]; and AirToxScreen 2019 ambient and exposure air toxics indices [16]. Release amounts were log-transformed (log1p); because Spearman uses ranks, this does not change the result. Each air toxics index is the county average, across its tract rows, of the summed concentrations of 177 pollutants (µg/m³), unweighted by toxicity.

Pope County

Pope and Hardin were compared descriptively with the median Illinois county, with 10 other southern Illinois counties (Alexander, Jackson, Williamson, Johnson, Gallatin, Saline, Pulaski, White, Union, Massac), and with all 100 other counties. Because Pope was selected for its high rate, Wilcoxon rank sum statistics are reported as descriptions only; the statewide scan accounts for that selection. Earlier IDPH county reports for 2008–2012 and 2013–2017 were checked for the same counties [20, 21].

1. National

Incomplete and inconsistent national data leave the true CHD burden uncertain. About 35% of U.S. births are not included in national data [1, 7], and totals for the same defects vary 13.6-fold between programs. Although CDC estimates about 40,000 CHD births each year [2], states track them separately and by different reporting criteria [3].

The highest and lowest totals for the same 16 defects in two different state programs are 417.6 per 10,000 in Kentucky and 30.7 in Maryland [1]. Kentucky counts defects found up to age 5 from hospital logs, clinics, and Medicaid records, while Maryland counts newborns only, mainly from hospital report forms. Kentucky records 322.3 atrial septal defects per 10,000; Maryland records 7.6 [1]. Infants with several defects are counted once per defect, so program totals are diagnoses, not infants, and cannot be read as prevalence.

Map of the 50 United States. 34 states are green or green-striped, meaning they are in NBDPN national CHD data for 2014 to 2018; Georgia and California are striped because only part of each state is covered; 16 states and the District of Columbia are grey. — Figure 1. National Birth Defects Prevention Network data, 2014–2018 [1]. Green: in national data. Striped: partial coverage. Grey: not in national data.

Table 1. CHD diagnoses per 10,000 live births by state program, 2014–2018 [1]
Rank	Program	Years	Live births	Per 10,000
1	Kentucky	2014–2018	275,341	417.6
2	New Mexico	2014–2018	114,267	374.3
3	Tennessee	2014–2018	405,499	298.6
4	Alaska	2014–2017	44,498	262.0
5	Mississippi	2014–2018	189,440	261.6
6	West Virginia	2014–2018	84,318	232.6
7	Virginia	2014–2018	503,868	212.7
8	Florida	2014–2018	1,107,528	211.8
9	New York	2014–2018	1,187,108	209.1
10	Vermont	2014–2018	28,878	200.1
11	Texas	2014–2017	1,581,796	199.6
12	Colorado	2014–2018	326,321	188.6
13	Ohio	2014–2018	692,695	186.0
14	Louisiana	2014–2017	238,758	178.5
15	Indiana	2014–2018	408,728	152.5
16	Arkansas	2014–2017	152,794	147.7
17	Delaware	2014–2018	54,474	146.5
18	North Carolina	2014–2018	601,595	141.0
19	Wisconsin	2015–2018	280,671	137.5
20	Minnesota	2014–2018	331,901	135.2
21	Oklahoma	2014–2018	259,040	133.6
22	Iowa	2014–2017	156,891	126.2
23	Illinois	2014–2018	765,308	123.0
24	New Jersey	2014–2018	509,721	117.0
25	Georgia (Metropolitan Atlanta only)	2014–2018	176,508	114.7
26	Utah	2014–2018	248,215	106.3
27	Massachusetts	2014–2018	354,475	97.0
28	Rhode Island	2014–2018	53,577	93.3
29	Kansas	2014–2018	181,997	81.7
30	South Carolina	2014–2018	286,801	73.4
31	Maine	2014–2017	48,796	65.9
32	California (selected counties only)	2014–2018	310,281	61.0
33	Arizona	2014–2018	418,278	32.8
34	Maryland	2014–2018	362,831	30.7

Seven programs reported 4 of the 5 years, so 65% is the share of U.S. births represented in these data. Programs also differ in which defects they report; Arizona, for example, reports no atrial or ventricular septal defects.

2. State

Illinois CHD rates cluster, and EPA-tracked exposures do not account for the pattern. Counties with similar CHD rates tend to sit next to each other: global Moran's I = 0.161, z = 2.84, one-sided p = 0.0023 (two-sided 0.0045) [9, 10]. With rook contiguity, I = 0.160, z = 2.79, p = 0.0027. None of the seven EPA measures, including toxic releases, Superfund sites, and air toxics, tracks against those rates.

Map of Illinois's 102 counties with one circle per county, sized by number of CHD diagnoses and colored by rate compared with the typical county. Purple marks LISA clusters, white outlines mark scan clusters, and Pope and Hardin at the southern tip are red. — Figure 2. Illinois county CHD rates, 2016–2020 [8]. Typical = the median Illinois county, 124 per 10,000 births. Much higher: above 186; higher: 136 to 186; typical or lower: 136 or less. Clusters: LISA [11] and scan test [12].

Table 2. Local Moran's I (LISA) classes, queen weights, p < 0.05 [11]
Class	Counties
High-High (high-rate cluster)	Brown, Sangamon
Low-Low (low-rate cluster)	Edwards, Hamilton, Pulaski, Wabash, Wayne, White, Williamson
High-Low (outlier)	Alexander, Franklin
Low-High (outlier)	None
Pope and Hardin	Not significant, using Illinois counties only. Both border Kentucky, where the fluorspar district continues [22]; with Kentucky county data, their neighbors and this result could change.

Table 3. Kulldorff Poisson spatial scan, significant clusters [12, 13]
Cluster	Observed	Expected	Relative risk	p
Winnebago	342	237.2	1.46	0.001
Adams, Brown, Calhoun, Cass, Greene, Hancock, Jersey, Macoupin, McDonough, Menard, Morgan, Pike, Sangamon, Schuyler, Scott	462	357.4	1.31	0.001
Iroquois, Kankakee, Will	694	601.0	1.17	0.030
Pope (not significant)	6	1.9	3.08	0.865
Hardin (not significant)	3	2.1	1.41	1.000

Table 4. How closely 7 EPA measures track against county CHD rates: Spearman correlation, n = 102 (0 = not at all) [14–16]
EPA measure	ρ	p
TRI facilities, 2016–2020	+0.100	0.315
TRI total releases (log)	+0.048	0.634
TRI water releases (log)	+0.053	0.598
TRI metal releases (log)	−0.018	0.856
Superfund NPL sites, all statuses	+0.128	0.199
AirToxScreen 2019, ambient	−0.053	0.600
AirToxScreen 2019, exposure	−0.054	0.589

This is a county-level result. It does not show that environmental exposures are unrelated to CHD, and none of the seven measures captures legacy mine waste in groundwater.

3. County

Pope County has the highest rate in Illinois, but small numbers and a state line keep it from being distinguished from chance. Pope's rate, 416.7 per 10,000, is the highest in Illinois but rests on 6 cases [8]. Pope ranks 1st and Hardin 11th of 102 counties; the two counties average 303.9 per 10,000, 2.5 times the median county (124.0) [8].

Map of southern Illinois and western Kentucky. Pope and Hardin Counties in Illinois are solid red. Crittenden, Livingston, and Caldwell Counties in Kentucky are red-striped. A dotted black line marks the state line along the Ohio River between them. — Figure 3. The fluorspar district spans the state line: about 7,400 residents in Illinois and 30,500 in Kentucky [4, 22, 23].

The district continues into Kentucky, outside Illinois data and under different reporting criteria [1, 22].
A statewide scan cannot see across the state line; within Illinois alone, Pope's rate is not significant (p = 0.865) [12].
Pope had 144 live births and Hardin 157 in 2016–2020: 301 births and 9 diagnoses in five years [8, 13]. If Pope's true risk were three times the state rate, a county of its size would show 6 or more diagnoses only 53% of the time (Poisson, expected 1.95).
Pope was selected for its high rate. Testing a county chosen for its high rate inflates false positives: in 2,000 simulated Illinois maps with no real differences, the highest county tested against the rest came out significant 57.5% of the time.
For description only: both belt counties rank above all 10 comparison counties (Wilcoxon W = 20 of 20, one-sided p = 0.021) and above most other Illinois counties (W = 191 of 200, p = 0.014). These p-values do not account for how the counties were chosen.

Table 5. The district and its Kentucky neighbors: borders and population, Census 2020 [22, 23]
County	Borders	In fluorspar district	Population, 2020
Pope, IL	Livingston, KY	Yes	3,763
Hardin, IL	Crittenden, Livingston, and Union, KY	Yes	3,649
Crittenden, KY	Hardin	Yes	8,990
Livingston, KY	Pope, Hardin	Yes	8,888
Caldwell, KY	Neither	Yes	12,649
Union, KY	Hardin	No	13,668

Illinois side: 7,412 people. Kentucky side: 30,527. Borders were measured from Census county boundaries. Kentucky county births and CHD counts were not obtained for this poster; CHD counts would require a request to the Kentucky Birth Surveillance Registry.

Table 6. Pope County CHD cases by report period [8, 20, 21]
Period	Pope	Hardin	Illinois	Case definition
2008–2012 [20]	0 (0.0)	0 (0.0)	87.2	Newborn hospital stay
2013–2017 [21]	0 (0.0)	3 (194.8)	112.9	Children under 2
2016–2020 [8]	6 (416.7)	3 (191.1)	135.4	Children under 2

Cells show diagnoses (rate per 10,000). Because 2013–2017 includes 2016 and 2017 with no Pope cases, Pope's 6 diagnoses fall in 2018–2020, assuming both reports count cases the same way.

Table 7. CHD diagnoses per 10,000 live births, Pope and Hardin and 10 other southern Illinois counties, 2016–2020 [8]
County	Diagnoses	Per 10,000	95% CI
Pope	6	416.7	152.9–906.9
Hardin	3	191.1	39.4–558.4
Alexander	4	125.4
Jackson	31	99.5
Williamson	26	71.5
Johnson	3	56.3
Gallatin	1	43.5
Saline	6	41.8
Pulaski	1	31.8
White	1	14.5
Union	1	11.4
Massac	0	0.0
Illinois median county		124.0
Median of the 10 comparison counties		42.6

Documented hazards

The district's ore carries lead, zinc, and barium minerals [4]. Minerva Mine #1 tailings in Hardin County contain barium, cadmium, and lead, though the 2007 health consultation found no apparent public health hazard from current exposures [5].
Neither county had a TRI-reporting facility in 2016–2020. Pope has had none since TRI began in 1987; Hardin's only filer, Prince Minerals Inc. in Rosiclare, reported lead releases from 1998 to 2006 [14]. Neither county has a Superfund NPL site under any status [15].
Pope's Millstone Public Water District draws groundwater from three wells and serves 5,565 people [18]. Its treated tap water, 2018–2023, measured arsenic 5.98 ppb, total trihalomethanes 57.0 ppb, fluoride 0.44 ppm, and barium 19.7 ppb, all below EPA limits [17]. These are treated tap water results, not raw groundwater [17]. Private wells are not regulated under the Safe Drinking Water Act [19].

Limitations

County-level (ecological), cross-sectional design; no individual exposures or confounders.
NBDPN and APORS count defects, not infants.
Illinois data only. Pope and Hardin border Kentucky, where the fluorspar district continues in Crittenden, Livingston, and Caldwell Counties [22], so LISA and the scan see only their Illinois neighbors. Kentucky counts defects up to age 5 [1], so its county data would need matching reporting criteria before being combined.
Pope was selected for its high rate, so tests on the same data are biased; the statewide scan and new data are fair tests of it.
Small counts give wide confidence intervals; Pope's rate rests on 6 diagnoses.
Water data come from one public district, after treatment, for 2018–2023.
The AirToxScreen index is an unweighted sum of concentrations, and the source files carry no pollutant names.

Conclusion, recommendations, and next steps

CHD surveillance has gaps at every level: national data are incomplete and inconsistent, Illinois clusters are not accounted for by EPA-tracked exposures, and the Fluorspar Belt is too small, and too divided by the state line, to test conclusively with Illinois data alone.

These results are hypothesis-generating and point to where surveillance and research should focus next.

Recommendations for best practice

Standardize CHD reporting criteria nationally.
Build data-sharing agreements between neighboring states.
Monitor legacy mining sites and private wells [19].

Next steps for this study

Retest Pope and Hardin with APORS data from 2021 onward, which played no part in selecting them.
Obtain Kentucky birth defects and birth data for the district (Crittenden, Livingston, and Caldwell Counties) [22], with reporting criteria matched to Illinois.
Sample Pope County water, including private wells, in NIU labs.

References

National Birth Defects Prevention Network. Major Birth Defects Data from Population-based Birth Defects Surveillance Programs in the United States, 2014–2018 (Birth Defects Data and Directory). NBDPN; 2022. nbdpn.org
Centers for Disease Control and Prevention. Congenital Heart Defects: Data and Statistics. cdc.gov. Accessed October 2026.
Centers for Disease Control and Prevention. Birth Defects Tracking. cdc.gov. Accessed October 2026.
Denny FB, Nelson WJ, Breeden JR, Lillie RC. Mines in the Illinois Portion of the Illinois-Kentucky Fluorspar District. Illinois State Geological Survey Circular 604; 2020.
Illinois Department of Public Health, under cooperative agreement with ATSDR. Health Consultation: Minerva Mine #1, Cave-in-Rock, Hardin County, Illinois. May 4, 2007.
Panno SV, Weibel CP, Li W. Karst Regions of Illinois. Illinois State Geological Survey Open File Series 1997-2.
Martin JA, Hamilton BE, Osterman MJK, Driscoll AK. Births: Final Data for 2018. National Vital Statistics Reports. 2019;68(13), Table 1. cdc.gov/nchs/data/nvsr/nvsr68/nvsr68_13-508.pdf
Illinois Department of Public Health. Birth Defects and Other Adverse Pregnancy Outcomes in Illinois, 2016–2020. Table 5. November 2023. dph.illinois.gov
Moran PAP. Notes on continuous stochastic phenomena. Biometrika. 1950;37(1-2):17–23.
Bivand RS, Pebesma E, Gómez-Rubio V. Applied Spatial Data Analysis with R. 2nd ed. Springer; 2013.
Anselin L. Local indicators of spatial association: LISA. Geographical Analysis. 1995;27(2):93–115. doi:10.1111/j.1538-4632.1995.tb00338.x
Kulldorff M. A spatial scan statistic. Communications in Statistics: Theory and Methods. 1997;26(6):1481–1496. doi:10.1080/03610929708831995
Illinois Department of Public Health. Number of Live Births by Resident County, Illinois Residents, 2010–2019 and 2020–2023. dph.illinois.gov
U.S. Environmental Protection Agency. Toxics Release Inventory, 1987–2023, facility reports compiled by the author.
U.S. Environmental Protection Agency. Superfund National Priorities List (NPL) Sites with Status Information. Downloaded October 2026.
U.S. Environmental Protection Agency. AirToxScreen 2019: Ambient and Exposure Concentrations by Census Tract, Illinois. Explanation of data
Environmental Working Group. Tap Water Database: Millstone Public Water District (IL1515050). Data 2018–2023. ewg.org. Accessed October 2026.
Illinois Environmental Protection Agency. Drinking Water Watch: Millstone PWD (IL1515050). Accessed October 2026.
U.S. Environmental Protection Agency. Private Drinking Water Wells. epa.gov/privatewells. Accessed October 2026.
Illinois Department of Public Health. Birth Defects and Other Adverse Pregnancy Outcomes in Illinois, 2008–2012: A Report on County-Specific Prevalence. Table 5. June 2019. dph.illinois.gov
Illinois Department of Public Health. Birth Defects in Illinois, 2013–2017, county-specific prevalence. Table 5. dph.illinois.gov
University of Kentucky, Kentucky Geological Survey. Geology of Kentucky: Vein Minerals. uky.edu
U.S. Census Bureau. QuickFacts: Pope and Hardin Counties, Illinois; Livingston, Crittenden, Caldwell, and Union Counties, Kentucky. Population, Census, April 1, 2020. census.gov/quickfacts. Accessed October 2026.

Data and reproducibility

The poster's statistics are recomputed from the source tables by one R script, reproduce_poster_numbers.R, which prints each value beside the poster value and marks it PASS or FAIL. Earlier versions of the script, run by the author in October 2026 (R 4.6.1; spdep, sf, tigris, smerc 1.8.6, readxl), passed every check. Values the script does not compute, such as the water results, TRI history, and Census populations, are traced to their sources in a separate provenance document. The script, the extracted source tables, and the provenance document are available from the authors.

Data sources: NBDPN Birth Defects Data and Directory 2022; IDPH APORS county reports (2008–2012, 2013–2017, 2016–2020); IDPH county birth counts; NCHS National Vital Statistics Reports; EPA TRI, Superfund NPL, and AirToxScreen; EWG Tap Water Database; Illinois EPA Drinking Water Watch; Census 2020 county boundaries.

Acknowledgements

Dr. Karen Samonds (Biological Sciences), and Prof. Michael Mutersbaugh (Mathematics), for teaching me calculus. You were right, math is fun.