Skip directly to search Skip directly to A to Z list Skip directly to site content




The Agency for Toxic Substances and Disease Registry (ATSDR) issued a public health assessmentdraft for public comment on October 5, 2000 for the Newtown Community in Gainesville, Georgia.Between October 5, 2000 and November 17, 2000 the public had the opportunity to providecomments on the draft public health assessment. During that time, ATSDR has addressed commentsand reader evaluations from residents, Georgia Department of Natural Resources, Division of AirQuality, and local officials. These comments/questions are presented below. Comments are printedin italics.

General Responses to Comments

ATSDR received several reader evaluations and two letters presenting formal comments from stateofficials and residents. An evaluation received from a concerned resident stated that the report wasnot comprehensive enough, that ATSDR should have performed more testing, and that ATSDRshould have suggested more testing that should be conducted. In addition, the resident stated that thesummary "gives the false impression that the testing was conclusive, showing no problems."

ATSDR attempted to make the public health assessment as comprehensive as possible. Every studythat has been conducted in the community was discussed in detail and analyzed as thoroughly as thedata would allow. ATSDR determined whether or not health problems should be expected based onthe analysis of these data.

The resident stated that ATSDR should have done more testing. ATSDR does not conductcomprehensive environmental sampling. ATSDR does from time to time collect a limited number ofsamples for screening purposes only. The health assessor was able to mobilize resources for one ofthese screening efforts, which is referred to in the document as an "exposure investigation".Comprehensive sampling is generally conducted by regulatory agencies such as those who havepreviously conducted sampling in Newtown.

Comment: The summary gives the false impression that testing was conclusive, showing noproblems...

ATSDR Response: The summary specifically explains the uncertainty and inconclusiveness of thedata:

"Some data may not be representative of potential exposures in this community"
"ATSDR was unable to determine the public health implications of particulate emissions because no data were available..."
"ATSDR was unable to assess health implications of past exposure because no historical environmental sampling data exists"
"no surface soil outside the playground has been investigated in the community"

Comments from Ga DNR, Division of Air Quality

Comment: Page 14, second paragraph and page H-2 should read "NorthEAST GeorgiaMedical Center" instead of NorthWEST....

ATSDR Response: Thank you, ATSDR made these corrections.

Comment: Page H-17 has an incomplete sentence about the risk of tobacco on liver cancer.."

ATSDR Response: Thank you, ATSDR corrected this section.

Comment: Because the data collected thus far do not suggest any reason to believe thatresidential soil would be unduly contaminated, Georgia EPD will have difficulty allocating ourlimited resources for sampling analysis of residential soils in the community.

ATSDR Response: ATSDR recommended surface soils be analyzed in the community outside theplayground because there is no characterization of residential soils for soils 0-3" in residential yards.However, ATSDR understands the need for justification to spend resources on additional sampling.ATSDR cannot recommend further sampling to EPA or Georgia EPD in the absence of anyindication of elevated contamination in soils. The Public Health Action Plan (p. 20) delineates theserecommendations to specific agencies. Please note the third bullet under "future activities", whichstates that residents are "pursuing grants to conduct surface soil (0-3") sampling in accordance withan ATSDR recommendation".

Comments from the Newtown Florist Club

Before these comments are addressed, please note that ATSDR site managers are "health assessors"and not "risk assessors". Health assessors examine public health concerns of a community andwhether or not it has been impacted by industrial and environmental contamination. Risk assessorscalculate the theoretical risk of disease based on risk assessment criteria and default exposurescenarios. The instances where residents mention "risk assessor" should be changed in these comments to [health assessor].

Regarding particulate dusts

Comment: There is no indication of analysis having been performed by the agency or any otherentity to determine the composition of the dusts. No particulate data was available for analysis atthe time of the report. No attempt was made to identify the potential for contamination from thispotential airborne source of contamination.

ATSDR Response: The metal concentrations in air collected by Ga DNR, Division of Air Qualitywere the result of analyzing total suspended particulate (TSP) data. The composition of the dusts arewhat are reported as metals concentrations in Ga DNR air data.

Page 20 of the public health assessment ("Public Health Action Plan"), second bullet under FutureActivities, states "....included in that report will be a particulate matter analysis from data providedby the Georgia Air Quality Branch". This refers to a future health consultation regarding air thatwill address historical air modeling and will also address particulate concentrations in air reportedby Georgia DNR. This report includes an in-depth analysis of the sources of all emissions within 4miles of the Newtown area. The specific contaminants were modeled to evaluate estimated exposureto these airborne contaminants. ATSDR expects to release this document in Spring 2001.

Comment: The fallacy of the argument that there is no known public health hazard is clearlybased on the rationales given by EPA for regulating TSP and PM2.5. There is in fact significantprobable health hazard due to exposures from Total Respirable Dusts in the community.

ATSDR Response: ATSDR did not have data at the time this report was written to comment onTSP concentrations. However, ATSDR has obtained data from Ga DNR since the release of thisreport and will comment on them in the ambient air public health consultation, which ATSDRexpects to release in Spring 2000. Ambient air data that were reviewed do not pose a public healthrisk to residents at the time data were collected.

Comment: While it may be justifiable to assume the deposited particulates were not toxic, thefact that the particulates were heavy enough to be deposited by Brownian motion indicated thatthere possibly existed some particulates that were respirable.

ATSDR did not make any assumptions about the toxicity of particulate or respirable dusts. ATSDRanalyzed the concentrations of contaminants in air that were reported, researched the scientificliterature and health-based exposure guidelines, and determined whether or not negative healthoutcomes could be expected based on data available to ATSDR at the time this report was written.

The Newtown Florist Club made the following recommendation: Review of the Title V air permits and TRI reports for current and previous years of the neighboring plants would be an important source of information regarding discharged pollutants. There was no mention of this as a source of air toxic exposure data. Thus the hazard identification for particulate fraction of the air quality is inadequate. Without an adequate hazard identification process reviewing all available facility information, exposure assessment is not possible and the following comment made on page 1 is invalid "Current ambient air contamination in the community is not a public health hazard".

ATSDR Response: As stated on pages 1, 19, and 20, air modeling has been conducted toreconstruct past emission concentration scenarios given specific meteorological conditions of thearea. This modeling used TRI data that exists in catalogued databases. In this case TRI data for allfacilities within a 4 mile radius has been pulled from a number of sources. Reviewing the Title Vpermits of facilities located in close proximity to the community may explain the types of emissionsa plant can release, to which media, and in what manner, but it would not help to determine whetheror not contamination exists in the community, which is the focus of this document. ATSDR is not anenforcement agency and does not regulate facilities under Title V. The particulate fraction of the airquality will be addressed in the same document as the air modeling, estimated to be released inSpring 2001.

ATSDR concludes that "based on all available data current ambient air contamination in thecommunity is not a public health hazard.....". There are no data which have been presented toATSDR which suggest that the lives of residents are in danger because of the levels ofcontamination detected in ambient air. No levels that have been reported represent concentrationsthat would be expected to result in negative health effects for residents.

Regarding section "Extent of Contamination-Air"

Comment: The state of Georgia has not justified why the monitors were located at this site,which is neither in the neighborhood, nor does it have any apparent meteorological relationshipto prevailing wind patterns.

ATSDR Response:
Georgia DNR, Air Quality Branch, has responded to this issue with the following comments:

"Several points must be considered before establishing an ambient air monitoring site, includingwind direction, purpose(s) of the study including scale of representativeness, and physical sitingcriteria and limitations. A meteorological station was established at the school prior to beginning airsampling. Wind roses produced from this data indicated that the wind direction was from theindustrialized area south of Newtown only 18% of the time and that the majority of the time, thewind direction was from some other direction other than the nearby industrialized area. Thedirection of the winds indicate that the majority of the time the source of air pollutants in theNewtown area are due to sources other than [these] facilities.

The Fair Street school site meets the siting criteria of representativeness for a middle scale study.Middle scale represents air quality areas on the order of about 100 meters to 0.5 kilometers andcharacterizes areas of several city blocks in size. The school also meets all of the physical sitingcriteria of unobstructed air flow, probe height, and availability. Several sites were investigated aspossible sites, including the playground. These sites did not meet physical siting criteria."

Ga DNR collected meteorological data for nine months prior to installing the air monitor. If the airblows form the southern industry 18% of the time, then Newtown residents are exposed to theemissions from those facilities 18% of the time, and the monitor is fairly accurately representingambient air conditions in the community. ATSDR is currently evaluating the appropriateness of theambient air monitoring location.

Residents offered the following recommendation: Relocate the air monitoring device to a location that has meteorological significance with respect to wind patterns in the community, and also add the capability of monitoring TSP/PM2.5. Evaluate this data weekly for VOCs, SOX, NOX, CO and particulate materials.

ATSDR Response: Ga DNR has a report that justifies the choice of location and a significantamount of meteorological data in their sampling protocol. This report is available from Ga DNR orATSDR if residents are interested in reviewing it. Ga DNR began sampling for PM2.5 at the FairStreet School location on February 14, 1999. This data was supplied to ATSDR for our review anduse and will be discussed in the air health consultation to be released in spring of this year. In orderto justify sampling for SOX, NOX, and CO, there would have to be a potential identifiable sourcefor these emissions in this community. These contaminants are generally related to combustionprocesses, such as power plants and boilers. However, these contaminants are ubiquitous in theenvironment, and can be released from cars and trucks, as well as gas stoves, fireplaces, and othersources inside the home. Because there are so many potential sources for the elevation of thesecontaminants in ambient air, it would be extremely difficult to identify any one source as a majorcontributor.

The Ga DNR collects data twice a month for the Fair Street monitor. Unless data existed whichindicated that VOCs were present in significantly elevated concentrations, the likelihood that DNRwould allocate limited time and resources to additional monitoring is small.

Regarding Chromium

The reviewer noted information regarding chromium on page 7, however it appears that these datawere not addressed in the Exposure Investigation, but rather in the Health Outcome Data section,which begins on page 13. We assume the reviewer is commenting on the hair analysis on page 15.

Comment: Since the type of chromium found in biological monitoring and environmentalsampling has not been completely characterized, this metal may pose a health hazard ifexposure is due to the more toxic form. The [risk] assessor assumes that the less toxic form ofchromium is the type of exposure that is assumed by the assessor.

ATSDR Response: Levels of chromium found in hair cannot be proven to be caused byenvironmental contamination. There is no environmental evidence which suggests that the levels ofchromium found in air and soil are elevated above normal levels. Therefore, it is not possible to linkthe elevated chromium in hair to exposure to chromium in air or soil. Chromium is a necessaryelement and can be found in vitamins and dietary supplements. It is also found naturally in food,water, and air.

The chromium type was not identified in reported environmental data (unspeciated), and thereforeATSDR assumed that the chromium detected in air, water, and soil was hexavalent chromium, themost toxic form. We feel this was the most protective basis on which to compare concentrations ofchromium in environmental media.

The community made the following recommendation: If there is no basis, then a risk assessment of chromium should be included or a criteria for exclusion should be given. This is a risk assessment and even if an environmental toxin is present in the environment and is found in biological monitoring, there is an assumption that there may be an association. Thus, it may be an exposure and possible route unless the assessor can justify why it is not. We need to see a justification or some type of risk assessment of the chemical, chromium.

ATSDR Response: ATSDR's justification of why no cancer risk assessment, from herein called acancer risk evaluation, for chromium was done is the following: chromium was not found to bepresent at elevated levels in a single environmental media (in air, water, or soil). ATSDR will notassume there is an association between a contaminant found in hair and the environment if there isno environmental evidence to support such an association. If there are no data which suggest thatchromium is present at harmful levels, conducting a cancer risk evaluation would not beappropriate. ATSDR determines the toxicological implications of exposure if there is a potential thatexposure has occurred and could result in adverse health effects. Contaminants of concern (COCs)are those for which there are no health based guidelines (or comparison value) or one which exceedsestablished health based guidelines. Chromium falls into neither of these categories.

Risk assessments are conducted by regulatory agencies as a risk management tool. ATSDRconducts health assessments and health consultations, which attempt to evaluate the public healthimplications of chemical contamination and exposure.

Regarding the Exposure investigation and benzene

Comment: The [risk] assessor does not indicate the accuracy, specificity, nor sensitivity of thetesting method, nor the EPA/NIOSH method number...without this kind of information [methodnumber] this data is not real science and cannot be used to assess the health of a community".

ATSDR Response: Any government agency collecting or analyzing data is required by law tofollow the appropriate applicable guidelines set forth by EPA. Data requirements for ATSDR healthassessments are specified at in the "Environmental DataNeeded for Public Health Assessments Guidance Manual". The specific method used to analyzebenzene and other organics is printed on the data results which were provided to residents shortlyafter ATSDR received lab results from the SUMMA canister analysis. The cover page of thoseresults have a section that reads "analysis method(s): TO-15". Specific information on this methodcan be found at This is a gas chromatograph/massspectometry analysis of toxic organics ("TO") EPA method 15.

The reference of this document is as follows:
EPA 625/R-96-010b
Compendium of Methods for the Determination of Toxic Organic Compounds in Ambient Air, Second Edition. January 1997. Compendium Methods TO-14A, 15,16,17. (NCEPI or AMTIC)

Comment: ensure the accuracy of the reported data regarding benzene, more than foursamples would have had to be taken."

ATSDR Response: Georgia DNR has been sampling the Newtown area with the Fair Streetmonitor for four years. ATSDR took a limited number of air canister samples to determine whetheror not they would be different than those collected by the air monitor, which they were not, with theexception of one sample (.5 miles outside the community). During the year of 1997, a total of 22samples were collected by Ga DNR and an additional 10 were collected by ATSDR. In total, therewere 32 samples taken that would detect the presence of benzene, and only one was detected at alevel that was markedly elevated. Even at that level, ATSDR determined that the concentrationdetected would not result in an increased risk of cancer. If requested, ATSDR is willing to continueto analyze additional years of data as they become available to monitor trends in emissionsconcentrations. However, at this time, no additional data are available.

Regarding Soils

Page 9, subsurface soils

Comment: It was indicated that only subsurface soil samples were taken, but earlier in thedocument it was indicated that surface soil samples were taken from the playground, (page 1-Soils section). The report is contradictory as to whether or not subsurface samples were taken inthe community.

ATSDR Response: The summary states that residential subsurface soils do not pose a public healththreat. It goes on to state that the surface soils collected in the playground do not pose a public healththreat, but that no additional surface soils have been collected outside the playground. There appearsto be some confusion over this issue, but the purpose of the summary is to give a concise summaryof the content of the report. ATSDR examined both surface and subsurface soil data, which arepresented in detail in the report. Surface soils as defined by ATSDR are soils collected between 0-3inches, and anything deeper is considered subsurface soil.

On page 8, there is a section about the soil borings collected by Ga DNR and University of Georgia.Subsurface soil borings were collected in twelve inch composite samples between 0 and 23 feet inthese studies. The discussion of surface soils begins on page 9, under the heading "ATSDRExposure Investigation". To alleviate some confusion, ATSDR will add the subheading "SurfaceSoil" to the heading "ATSDR Exposure Investigation". ATSDR was the only agency to collect andanalyze surface soils and these samples were only taken in the playground. Five samples were takenin the playground, and although ATSDR does not recommend further sampling in the playground,there is a lack of surface soil data (0-3") in residential yards. This is the data gap that ATSDR refersto when it recommends additional surface soil collection and analysis in residential yards (p.1, p.19,p.20).

Comment: In addition, most [risk] assessors are aware that all humans ingest some amount ofsoil; more importantly, until 7 years ago, Newtown residents were eating food grown in theirhome gardens. As a result, soil characterization is very important to model possible exposuresdue to soils ingestion, food chain contamination, and most importantly to obtain an exposureassessment from surface soil. We also agree with the conclusion that subsurface soilcontamination does not pose a current health hazard to residents...

ATSDR Response: ATSDR is aware that soils are inhaled and ingested by adults, and to a greaterextent, children. ATSDR compares levels found in soils with health based guidelines for soilingestion and inhalation. Unfortunately, there are no biota data to analyze in order to determinewhether or not vegetables grown by residents were in fact contaminated. Unless residents areexcavating soils for construction or some other purpose, it is unlikely that they would be exposed tosubsurface soil contamination. However, even if they were exposed to the levels detectedunderground in this area, adverse health effects are not likely. ATSDR has stated that additionalsurface soil data are needed to fully characterized soil contamination and potential exposure (p.1).

Comment: If surface soil samples were taken, why were neither the number of samples, noranalysis information presented in the report? Were the samples representative? Did it indicatestatistical significance for any contaminant?

ATSDR Response: On page 6, where the exposure investigation was first mentioned, surface soil sampling is explained briefly:

In January, 2000 ATSDR conducted an exposure investigation, composed of five compositesurface soil samples in the community playground and ten air canister samples. Soils weretested for metals, and air was tested for VOCs....Exposure investigations are not in depthcollections of environmental data, and do not provide complete characterization of sampled media because of the small number of samples collected.

ATSDR agrees that mentioning the soil samples here is confusing and inappropriate and will add the specific soil sample information to the aforementioned section that was titled "ATSDR Exposure Investigation" on page 9. The data for the surface soil analysis is in fact present in the document in Appendix E, page 7. The discussion on page 9 under "ATSDR Exposure Investigation" analyzes these results. Because of the small number of samples taken, the calculation of statistical significance is not an appropriate method for analyzing surface soil data.

Residents recommended a section on exposure to contaminants through the food chain. At the levelsdetected when soil samples were taken, samples evaluated by ATSDR would not be of concernthrough the uptake of garden vegetables. However, no specific garden soil sampling data areavailable to evaluate the likelihood that residents were exposed to contaminated soils through thevegetables they ate.

Regarding health outcome data

Page 12

Comment: In addition to the studies listed here in this section, other pertinent data (with respectto the lupus study by Emory University) that could be used to analyze the existing health datawith respect to environmental exposures include:

Kilburn, K.H., Warsha, R.H., Prevalence of symptoms of systemic lupus erythematosus (SLE) andof fluorescent antinuclear antibodies associated with chronic exposureto trichloroethylene andother chemicals in well water. USC School of Medicine.

Conrad, K. et al. Systemic Lupus Erythematosus after heavy exposure to quartz dust in uraniummines: clinical and serological characteristics, Lupus 5(1): 62-9, 1996, Feb.

ATSDR Response: While these studies may help to substantiate a growing body of evidence ofassociations between environmental exposures and lupus, they are not pertinent to this community.Residents are not exposed to contaminated well water, nor are they exposed to high levels of quartzdust in uranium mines. ATSDR acknowledges that although no specific associations have beenstrongly substantiated, it is possible with the chemical triggers of lupus in medications, that theremay also be environmental triggers through toxins. ATSDR found numerous studies investigatingexposure to silica dust and lupus, however could not find any additional studies documenting theassociation between other contaminants and lupus. Many studies are necessary to make scientificfindings credible. There are a large number of studies available on the subject of exposure to silicadusts and scleroderma and lupus incidence, but inconclusive findings regarding the associationbetween other contaminants and lupus. Silica exposure is highly unlikely in this community, unlessresidents are occupationally exposed. ATSDR is not aware of any facilities in the Newtown areawhich uses or handles silica or silica dusts.

Comment: The [risk] assessor added sections in the appendices relevant to cancer prevalence.It would make this information clearer if the [risk] assessor would list cancers found in theNewtown community (prevalence ratios) and compare this to prevalence ratios for anotherpopulation or with national figures in summary tables side by be relevant to thispopulation, the information presented should be ...for African-Americans...

ATSDR Response: We agree that this data would be useful to residents to illustrate ourconclusions about cancer incidence in the Newtown community. Below is the information residentsand reviewers have requested:

Residents in this area are concerned that emissions from neighboring facilities may be causingexcess cancer in their community. In response to these concerns, ATSDR investigated cancerstatistics for the Newtown community to determine whether or not cancer rates in the community arestatistically different compared to United States rates for African Americans. No formal report ofthis comparison was used to interpret the comparison, and ATSDR analyzed raw data generated bythe Northeast Georgia Medical Center.

Almost all diseases or health outcomes occur at different rates in different age groups. Most chronicdiseases, including most cancers, occur more often among older people while other outcomes, suchas many types of injuries, occur more often among younger people. Therefore, the most commonhealth problems in a community will be influenced by the age distribution within the community.This community has a very large elderly population (22% are 65 or older), and so we expect to see ahigher number of cancers diagnosed than in a younger population.

One means of comparing the pattern of health outcomes in communities of different sizes is tocalculate an incidence or mortality rate, which is the number of new cases or deaths divided by thesize of the population. In chronic diseases and injuries, rates are usually expressed in terms of thenumber of new cases or deaths per 100,000 people per year. Adjusting rates for age allows for directcomparison between populations with potentially different age distributions. The cancer ratesdiscussed below are age adjusted cancer rates.

In this analysis, cancer incidence rates of residents living in the Newtown community werecompared to cancer rates of the US population. Newtown cancer incidence rates were derived fromthe Northeast Georgia Medical Center and population estimates from the 1990 census.

US cancer incidence rates were extrapolated from the Surveillance, Epidemiology, and End Resultsprogram (SEER) of the National Cancer Institute. The SEER database tracks cancers in five states(Connecticut, Hawaii, Iowa, New Mexico, Utah) and six metropolitan areas (Atlanta, Detroit, LosAngeles, Seattle/Puget Sound, San Francisco/Oakland, San Jose/ Monterey). With respect toselected demographic and epidemiologic factors, these areas are reasonably representative subsets ofthe United States population. The disease rates and patterns documented in the SEER database areaccepted as fairly accurate representations of the disease incidence rates and patterns of the UnitedStates as a whole. Therefore, the 'normal rates' of disease are often based on cancer rates in theSEER areas.

The cancer incidence and mortality data suggest that age-adjusted cancer rates for all cancers areactually lower in Newtown than in the United States (400.23/100,000 vs. 482.25/100,000).Newtown rates of cancer of the bladder, cervix, esophagus, kidney, liver, and for leukemia slightlyexceeded those of the US population. However, the differences between the rates were not found tobe statistically significant. Please see the "technical notes" section following the cancer incidencetables for the methods ATSDR used to make this determination.

Presented on the next two pages are the age-adjusted incidence of cases per year for the NewtownCommunity when compared to the age-adjusted incidence of SEER sites representing the U.S.population. ATSDR used the incidence for African-Americans in the SEER database based on theracial makeup of the Newtown community. The cancers listed were those diagnosed during aseventeen year study period in the Newtown community.

Table 1. Rates of Cancer in the Newtown community and United States SEER sites 1980-1996

Newtown cancer rates compared to national SEER rates for African-Americans

Cancer type

Actual numberof Cancers

age-adjusted cancer rate

U.S. Standard (SEER)
age-adjusted cancer rate

All Cancers36400.23/100,000482.25/100,000
Bladder3 Male 21.6/100,000*
Female 10.82/100,000*
Male 15.83/100,000
Female 5.75/100,000
Breast 443.29/100,00092.9/100,000
Cervical4 <50 years old 43.29/100,000*
50+ N/A
<50 years old 7.45/100,000
50+ years old 36.55/100,000
Colorectal5 Male 21.65/100,000
Female 32.47/100,000
Male 59.5/100,000
Female 46.39/100,000
Esophageal2 Male 10.82/100,000
Female 10.82/100,000*
Males 17.15/100,000
Females 4.7/100,000
Kidney1Male 10.82/100,000*Male 4.67/100,000
Leukemia3 Male10.82/100,000
Female 21.64/100,000*
Male 11.9/100,000
Female 7.38/100,000
Liver1Female 10.82/100,000*Female 2.29/100,000
Lung 1 (in Newtown)Male 10.82/100,000Male 124.1/100,000
Oral/pharyngeal3 Male 21.6/100,000
Female 10.8/100,000*
Male 23.7/100,000
Female 7/100,000
Pancreatic 2 Male 10.82/100,000
Female 10.82/100,000
Male 16.4/100,000
Female 12.3/100,000
Testicular1<50 years old 10.82/100,000< 50 years old .95/100,000
Uterine150+ years old 10.8/100,00050+ years old 58.02/100,000
* these rates exceeded SEER rates and therefore were chosen to be tested for statistical significance

Table 2. Statistical test of significance for differences in Newtown rates and SEER rates

95 % Confidence intervals for differences in cancer incidence rates between 1980 and 1996
Cancer type Standard incidence Ratio (SIR) Confidence interval Statistically significant?
Bladder Male 1.36
Female 1.88
Male .034-7.57
Female .0475-10.47
Cervical 5.81.1469-32.36No
EsophagealFemale 2.3.058-12.8No
KidneyMale 2.32.059-12.9No
LeukemiaFemale 2.9.073-16.2No
LiverFemale 4.7.12-26.18No
Oral/pharyngealFemale 1.54.039-8.58No

Technical Notes: Methods for Analyzing and Interpreting Cancer Incidence Data

In order to be able to analyze and interpret cancer incidence data, it is necessary to convert thenumber of cases we observe to ratios. Using ratios allows us to compare the number of cases in thepopulation living in Newtown and the area of interest with a reference population to determine ifthere is an excess of a particular disease of health condition. When interpreting cancer data, anobserved occurrence is compared to an "expected" occurrence using ratios. The expected occurrenceis based on the occurrence observed in a reference population, in this case the SEER population forAfrican Americans. For cancer, the ratio of observed to expected number of cases (incidence) wasexamined, and the information was further standardized to eliminate possible effects due to age.These ratios are referred to as the standard incidence ratio (SIR).

Specifically, the SIR is the observed number of cases divided by the expected number of cases. Aratio of 1.0 means that the number of cases observed in the population being evaluated is equal tothe number of cases expected based on the rate of disease in the reference population. A ratio greaterthan 1.0 indicates that more cases occurred than expected; and a ratio of less than 1.0 indicates thatfewer cases occurred than expected. Accordingly, a ratio of 1.5 is interpreted as 1.5 times as manycases as the expected number, and a ratio of 0.9 indicates 0.1 fewer cases than would be expected.

Caution should be exercised, however, when interpreting these ratios. The interpretation of a ratiodepends on both the size of the ratio and the number of cases used to calculate the ratio. Forexample, a ratio of 1.5 based on two expected cases and three observed cases indicates a 1.5 timesexcess in cancer, but the excess is actually only one case. Conversely, a ratio based of 1.5 on 200expected cases and 300 observed cases represents the same 1.5 times excess in cancer , but becausethe ratio is based on a greater number of cases, the estimate is more stable. It is very unlikely that100 excess cases of cancer would occur by chance alone.

Experience has shown us that we can expect a certain amount of chance variation when looking atthe occurrence of different health conditions. Statisticians have developed methods to take this intoaccount. One method is to calculate a confidence interval for the SIR. A 95% confidence interval(CI) is calculated to determine if the observed number of cases is significantly different than theexpected number, or if the differences may be due solely to chance. A 95% CI is the range ofestimated values that has a 95% probability of including the true ratio for the population. Theconfidence interval is a statistical measure of the precision of the risk estimate. If the confidenceinterval contains 1.0, no statistically significant excess of disease is indicated.


Standardized (Adjusted) Rates: help control for demographic differences between populations beingcompared. Adjusted incidence rates estimate what the incidence rates for populations would be iftheir composition were similar ro that of a comparison, or standard, population (and, therefore, toeach other). Adjustments can be made for various characteristics that influence incidence rates,including their age, race or ethnicity, and gender.

Although a crude rate (non-adjusted rate) is a valuable summary measure, comparison of crude ratesbetween populations can be problematic if demographic characteristics (such as age distribution)that affect health outcomes differ between the populations. The overall crude incidence rate for apopulation depends on not only the incidence rate for each age group, but also the proportion ofpeople in each age group.

Age-adjustment helps control for differences in the age structure of the population. Age-adjustedincidence rates for two populations are calculated by multiplying the age-specific incidence rates foreach age group by the proportion of people in the same age group in the standard population. Thesum of these products is the age-adjusted, or age-standardized, incidence rate for each of thepopulations.

Statistically significant: there is less than a certain percent change (usually selected at 5%) that theobserved difference is merely the result of random fluctuation in the number of observed cancercases. For example, if the confidence interval does not include 1.0 and the interval is below 1.0, thenumber of cases is significantly lower than expected. Similarly, if a confidence interval does notinclude 1.0 and the interval is above 1.0, then there is a significant excess in the number of cases. Ifthe confidence interval includes 1.0, than the true ratio may be 1.0, and it cannot be concluded withsufficient confidence that the observed number of cases reflects a real excess or deficit. As long asthe 95% confidence interval contains 1.0, that indicates that the ratio is still within the range onemight expect based on the disease experience of the comparison population.

In addition to the number of cases, the width of the confidence interval also reflects the stability ofthe ratio estimates. For example, a narrow confidence interval (e.g. 1.03-1.15) allows a fair level ofcertainty that the calculated ratio is close to the true ratio for the population. A wide interval (e.g.0.85-4.50) leaves considerable doubt to the true ratio, which could be much lower than or muchhigher than the calculated ratio.

No Newtown cancer rates were found to be significantly higher than those of the referencepopulation (SEER). The cancer rates that had an SIR of 1.0 or greater were tested for significance(table 2). The confidence intervals for these cancers were very wide, and all included 1.0. As theexplanation above states, if a confidence interval includes 1.0, it is not considered statisticallysignificant. A wide confidence interval usually means that the sample size is too small, there weren't enough cases for comparison, or the population was too small to calculate significant statistics.

Table of Contents The U.S. Government's Official Web PortalDepartment of Health and Human Services
Agency for Toxic Substances and Disease Registry, 4770 Buford Hwy NE, Atlanta, GA 30341
Contact CDC: 800-232-4636 / TTY: 888-232-6348

A-Z Index

  1. A
  2. B
  3. C
  4. D
  5. E
  6. F
  7. G
  8. H
  9. I
  10. J
  11. K
  12. L
  13. M
  14. N
  15. O
  16. P
  17. Q
  18. R
  19. S
  20. T
  21. U
  22. V
  23. W
  24. X
  25. Y
  26. Z
  27. #