MONICA BROWN, PHDCANCER EPIDEMIOLOGIST

THE CALIFORNIA CANCER REGISTRY

Cancer Concerns

CANCER EPIDEMIOLOGY (EPI-272)JANUARY 21, 2011

UCD, DEPT OF PUBLIC HEALTH

What Drives the Public’s Concern of the Clustering of Cancers in Communities and

Workplaces?

There is considerable public concern that environmental exposures cause an excess in cancers in some communities.

The public believes environmental pollutants/toxins increase risk of cancer - although, there’s no evidence that there is increased risk to the general population in amounts that are typically present in the air, soil or drinking water.

Cancer clusters may be suspected when people notice that several family members, friends, neighbors or co-workers have been diagnosed with cancer, when the distribution of cancers may be “normal” given the age, sex, race/ethnic and lifestyle of that group.

… continued

Other phenomena that may drive suspicion of environmental cancer clusters are... Media reports sensationalized cancer clusters Distrust of government, manufacturing and

business Fear that we’ve created an environment filled with

hazards that is causes us and our families harm The perceived inability to control cancer risk and

environmental hazards Ever changing and varied Public Health (PH)

messages

What We know

Cancers are common!Cancer incidence varies by age, sex, race/ethnicity

& risk factorsCancers are complex diseases - PH has

oversimplified cancer Use of the singular – “cancer” instead of “cancers” Lumping all non-clinical risk factors as “environmental”

Communities members are often similar - age, SES, race/ethnicity & lifestyles – these factors contribute more to cancer incidence than shared environment

Knowledge of cancer causes, its distribution and prevention varies greatly in the general public – PH has done a poor job educating the public about cancer;

therefore the public has many misconceptions about cancer & cancer clusters

Common Misconceptions about Cancer

Cancer is uncommonCancer is too common - 1 in 5 Californians will have a cancer in their lifetime

Cancer is one disease Cancer is a general term for many diseases – most with different etiologies

Scientist understand cancer development

The etiology of many cancers is unknown

Young people shouldn't get cancer

Cancers can be found in every age group

All cancers can be caused by a single factor

Cancers are thought to be caused by a combination of factors

Exposure to a known carcinogen and the onset of cancer is certain and immediate

Exposure to a carcinogen and the onset of cancer is not certain, other factors, some known, may be required. When cancer does develop, the onset can be decades from exposure

Environmental exposures are the primary cause of cancer

Lifestyle has been associated with 68% of the most prevalent cancers

Causes of Cancer

Lifestyle68%

Environment/Occupation

19%

Family History/Genetics

13%

Lifestyle•Tobacco Use 30%•Diet 10%•Physical Inactivity 5%•Alcohol Use 3%•Other 20%

Environmental/Occupation•Occupation 5%•Viruses/other biologics 5%•SES 3%•Pollution 2%•Radiation 2%•Other 2%

Family History/Genetics•Family History 5%•Prenatal Factors/Growth 5%•Reproductive Factors 3%

Source: Harvard Report on Cancer Prevention, 1996

0

10

20

30

40

50

60

70

80

90

<5 5-9 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85+

Age at Diagnosis

ASIR

per

100

,000

Female Breast Prostate Female Lung Male Lung Female Colorectal Male Colorectal

Age-Specific Incidence Rate (ASIR) for most prevalent cancers

Common Misconceptions about Cancer Clusters

Clustering is uncommon

Clustering of health events is common - some random (1%) some not. Shared social-demographic characteristics and/or similar lifestyles explains some health event clustering.

Several cancer cases make a single cause cluster

We expect a certain number & certain types of cancers in every neighborhood/workplace.

If there are several cases of cancers in a community, of different types, they must come from the same source.

If there are several different types of cancers in a community, there are likely several different causes.

Examples of Documented Cancer Clusters

Cluster Characteristic AgentRare

Incr dz freqAngiosarcoma of the liver among factory workers Rare

Vinyl Chloride

Mesothelioma among Native American silversmiths Rare asbestos

Incr dz frequnique pop

Diethyl-stilbestrol

Adenocarcinoma of vagina among young women

Kaposis sarcoma among young men HIV

The California Cancer Registry

The California Cancer Registry (CCR) is administered by the California Department of Public Health (CDPH).

The CCR is a true population-based registry. Cancer reporting is mandated for hospitals and

physicians.Data collected by the registry are used:

To monitor incidence and mortality. For research into the causes, cures and prevention of

cancer; To produce reports including the state and regional

annual reports and the American Cancer Society’s Cancer Facts and Figures; and

The evaluation of community cancer concerns.

When a Californian has a Cancer Concern: the Role of the CCR

The CCR and it’s regional cancer registries respond to numerous requests for evaluation of community and workplace cancer concerns.

The registry’s role is to statistically assess whether the number of cases of targeted cancers observed in a community or workplace are significantly greater than what would be expected.

If there is a statistically significant excess of cases, report to the Environmental Health Investigations Branch (EHIB) of CDPH who will investigate.

The Role of the CCR, continued

The CCR does not: Conduct epidemiologic “outbreak”, clinical or

laboratory investigations. On-site surveys of residents or employees to

assess risk. Direct others in exposure assessments. Coordinate the efforts of other state and county

agencies in their investigations.

AN AGGREGATION OF CANCER CASES THAT HAS BEEN DETERMINED TO BE UNUSUAL WHEN

COMPARED TO THE CANCERS THAT WOULD BE EXPECTED IF THE GROUP OF LOCATION IN QUESTION HAD THE SAME CANCER RATES AS THE UNDERLYING

POPULATION.

THE CLUSTER MUST DIFFER SUBSTANTIALLY FROM THE EXPECTED PATTERN IN NUMBER, TYPE, OR THE

AGE OF CASES.

We define a Cancer Cluster as…

AN UNUSUAL AGGREGATION, REAL OR PERCEIVED, OF HEALTH EVENTS THAT ARE

GROUPED TOGETHER IN TIME OR SPACE AND THAT ARE REPORTED TO A HEALTH AGENCY*

The CDC defines a Cluster as …

* Guidelines for Investigating Clusters of Health Events, 1990, Centers for Disease Control and Prevention

Procedures

1. Obtain Information from Informant2. Provide Cancer Education and Information 3. Assess Cancer Concern - Determine if

Further Analysis is Needed4. Explain Procedures, Limitations of Methods

and Provide a Timeline5. Consult and Notify Relevant Officials6. Perform Assessment7. Communicate the Results of Assessment

Step One: Obtain Information

Caller’s name & address; affiliation (community member)

Number of specific cases observed

Cancer type(s) observedAge, sex, race/ethnicity of casesGeographic area or groupTime period of concernMethod of observation – how

did the caller learn of the cases

Step Two: Provide Cancer Education & Information

Education The frequency of specified cancers in

their community or County Risk factors for specified cancers If knowledgably, discuss agent and/or

exposure Information

American Cancer Society (ACS) Centers for Disease Control and

Prevention (CDC) The National Institutes of Health

(NIH) Agency for Toxic Substances and

Disease Registry (ATSDR)Note: Do not assume that everyone has access

to or can use the internet

Indications for Statistical Evaluation Other Considerations Are cancers unusual in

number, type or age of patients?

Has a potential carcinogenic agent been identified? If a specific exposure is

suspected – test 1st – call County Environmental Health, Environmental Protection Agency (EPA) or if workplace, Occupational Safety and Health Agency (OSHA)

Is there a plausible exposure pathway?

Is the request coming from a another State agency or from a County Health Department?

Is informant representing a community or workplace action group?

Are children involved? Is this perceived cancer

cluster “political” or is it already being followed by the press?

Step Three: Determine if Further Analysis is Needed

Step Three: Determine if Further Analysis is Needed

Unusual Cancers

IncreasedCancer

Frequency

Occurrencein UniquePopulation

or or

+CCR

Documented

Cases

BiologicPlausibility

Further Action Warranted

+Carcinogenic Agent

Step Four: Explain Procedure, Limitations & Provide Timeline

Procedure We use registry data to confirm case

information & determine clinical characteristics of cancers

We use census data for denominators (population at risk)

Perform calculations, write report to county & state.

In the event of a statistically significant excess of cancers, we refer case to EHIB for investigation

Limitations CCR will not contain most recently

diagnosed cases Only a substantial increase in risk is likely

to be detected We lack information on length of

residence and risk factors that may contribute to developing cancer

Timeline 1-3 months

Workplace Cancer Concerns: Barriers to Evaluations

Obtaining appropriate information on ill & well (population at risk) employees from employers is difficult to impossible.

If necessary, must obtain permission from employees to access their medical records.

Assessing biologic plausibility: Is the suspected agent at work associated with increased risk of reported cancers? Does workplace exposure have an impact? direct vs. indirect; length

of exposure (workday/year(s)); mode of exposure (eat/drink, inhale etc.)

What other risk factors could increase risk of developing reported cancers – smoking, drinking & diet – that cannot be assessed?

Separating endemic cancers from those reported: what cancers would be “normal” for this employee group.

Are there behaviors that are common in this employee group?

Step Five: Consult and Notify Relevant Officials of Report

Management hierarchy of CDPH

County Health OfficerWorkplace management

Step Six: Perform Assessment

Define geographic area by census tract

Review observed casesGenerate expected number of

cancers 5-year type-, age-, sex- and race-specific

rates for the state or region U.S. Census Bureau year 2000

population data for the census tract(s)Compare cases observed and

expected, calculate 99% confidence interval

Determine whether a statistically significant excess is found

Step Seven: Communicate Results

Write letter or report describing concern and results of assessment to the … Informant County Health Officer CSRB management hierarchy

If results show a statistically significant excess in cases, include … EHIB CDPH public affairs office

Challenges in Communicating Results

Science Scientific evidence is inconclusive, contradictory and ever-

changing Current scientific evidence is not absolute. Therefore, we

cannot give definitive answers. Scientific method - descriptions of methodological limitations

and results can sound evasive.Complicated scientific Concepts:

Random events 1% of all census tracts would have higher or lower cancer rates

simply by chance No one has ever called me and said “… there’s too few cancers

in my neighborhood”. public seemingly can only grasp concept if discussing the lottery.

… continued

Epi & Stat Concepts Often case and/or population numbers are too small for

appropriate statistical analysis, and we are unable to conduct analysis. sometimes viewed as demeaning the current number of cases. sometimes viewed as evasive or manipulative.

For environmentally based cancer concerns, we examine only related cancers not “all cancers” due to etiologic differences in cancers – often public thinks all cancers are germane.

Causality - if cluster confirmed statistically, doesn’t mean cancer is due to a single causal pathway.

Epidemiologists & Statisticians (us) Objectiveness viewed as lack of empathy. Expertise viewed as “Ivory Tower’ism” We are not good at saying we don’t know

Rules

1. Accept and involve the public as a partner.

2. Plan carefully and evaluate your efforts.

3. Listen to the public's specific concerns.

4. Be honest, frank, and open.5. Work with other credible

sources.6. Meet the needs of the

media.7. Speak clearly and with

compassion.

Your primary goal is to produce an informed public, not to defuse public concerns.

Different goals, audiences, and media require different actions.

People often care more about trust, credibility, competence, fairness, and empathy than statistics and details.

Trust and credibility are difficult to obtain; once lost, they are almost impossible to regain.

Conflicts and disagreements among organizations make communication with the public more difficult.

The media are usually more interested in politics than risk, simplicity than complexity, danger than safety.

Always acknowledge the tragedy of an illness or death. People can understand risk information, but they may still not agree with you; some people will not be satisfied.

The Seven Cardinal Rules of Risk Communication

From: The Seven Cardinal Rules of Risk Communications, Covello and Allen 1988

In Conclusion

Cancer clusters DO occur in communities, but are difficult to investigate and nearly impossible to prove. Our tools to investigate are crude and we often lack

pertinent information or time to see the natural history of events.

Cancer never 1st disease manifestation in true cluster From exposure to diagnosis can be 20 – 50 years, depending

on carcinogen Most prevalent cancers are not strictly caused by

environmental exposures – i.e., lung or prostate cancer Ignorance: what we think is harmless today, tomorrow we

may learn is dangerous.We must take responsibility and precautions to

safeguard our health.

For More Information on Cancer Clusters

ACS: http://www.cancer.org/Cancer/CancerCauses/OtherCarcinogens/GeneralInformationaboutCarcinogens/cancer-clusters

NIH: http://www.cancer.gov/cancertopics/factsheet/Risk/clusters

CDC: http://www.cdc.gov/nceh/clusters/ ATSDR:

http://www.atsdr.cdc.gov/csem/cluster/docs/clusters.pdf

Harvard School of Public Health, Disease Risk Profile: http://www.diseaseriskindex.harvard.edu/update/hccpquiz.pl?lang=english&func=home&page=cancer_index

Thank You!