Maine Public Health Data Reports

Technical Notes

Term Description
Age-adjusted Rates Defined What are age-adjusted rates?
Age-adjusted Rates Why age adjust? And how is age-adjustment done?
Association and Causation Why association does not necessarily mean causation?
Confidence Intervals The true value of a statistic.
Data Variability Issues around data variability and small numbers.
Defining Disability Defining disability.
Measuring Disability How disability is measured.
Five-year Trailing Average Combining five years of data.
Incidence and Prevalence Two ways to measure disease rates.
Making Two Populations Comparable There are two ways to make two populations comparable when known characteristics are distributed differently between them.
Margin of Error The degree of uncertainty in an estimate.
Medium and Means Two ways to measure central tendency.
Quality of Life Measuring quality of life using healthy days.
Race and Ethnicity How race and ethnicity is asked.
Ratio, Proportion, Rate What constitutes and numerator and the denominator.
Sensitivity and Specificity Efficacy of screening tests.
Sexual Orientation How sexual orientation is asked.
Tobacco Use Measuring tobacco use.

 

Age-Adjusted Rates

What are Age-Adjusted Rates?
Age-adjustment is a method used to better ensure comparability of estimates (e.g. rates) with respect to age. The age distribution of a population may change over time and differ from place to place. Because some health conditions or diseases are more common in certain age groups of people, it can be misleading to compare rates or prevalence estimates of populations if the age distribution of the populations compared are different. Apply the U.S. 2000 Census standard population to Maine's age-specific rates. Age-adjusted rates are relative and should not be considered exact rates that necessarily represent the true underlying burden of disease in the population.

Why Age-Adjust?
Different communities have different age structures. The age structure of a community determines what kind of health problems will be more common. For example, a community made up of more families with young children may have more bicycle accidents than a community with more individuals who are older. Likewise, a community with more individuals who are older will have more chronic disease, including cancer. Age adjustment allows rates of disease to be compared between different communities with different age structures.

 

[Back to Top]

Association and Causation

Why Association Does Not Necessarily Mean Causation?
Epidemiologists often look at associations of events and diseases, for instance, exposure to cigarette smoking and developing lung cancer. Although studies showing associations are often reported in medical and public health journals and subsequently picked up by the media, an association does not necessarily imply causation.

Why Age-Adjust?
Difference communities have different age structures. The age structure of a community determines what kind of health problems will be more common. A community made up of more families with young children will have more bicycle accidents that a community with more individuals who are older. Likewise, a community with more individuals who are older will have more chronic disease, including cancer. Age adjustment allows rates of disease to be compared between different communities with different age structures.

Four common possibilities that can explain an association:

  1. The association can be due to chance. Tests of statistical significance are important in determining the probability the association is due to chance. Some examples include a T-test (which compares the means of two populations) and a chi-square test (which also compares the outcomes of two populations). P values are often measured from tests of statistical significance in order to assess the probability a test result occurred by chance. By convention, if the P value is less than or equal to 0.05, there is no more than a 5% or 1 in 20 probability the result is seen by chance and, therefore, the association is probably statistically significant. Even if an association is true and due to an effect causing a disease, the P value can be large because of a small sample size. Confidence intervals are used to show the range of data results within which the true values are likely to be. Generally, the width of the confidence interval is affected by the sample size - a larger sample size results in a narrower confidence interval.
  2. The association can be due to a bias such as when non-comparable criteria are used to enroll participants (selection bias), or when non-comparable information is obtained from the different populations studied (observation bias), or when investigators elicit or interpret information differently (interviewer bias), or the participants report events in a non-comparable manner (recall bias).
  3. The association can be due to a mixing of effects between the exposure, the disease, and a confounding factor - a third factor that is associated with the exposure and that can affect the risk of developing the disease. Age is a common confounding factor, especially with many chronic diseases. Therefore, when comparing chronic disease rates between different time frames or geographical areas, the rates should be age-adjusted in order to make them comparable with respect to age. Alternatively, disease rates only for pertinent age groups should be compared.

The event (exposure) may contribute toward causing the disease; i.e., the association is a causal one. Determining this requires addressing all of the above issues and also looking at the strength or magnitude of the association, the biologic credibility, consistency with other results, if the time sequence makes sense, and if there appears to be a dose-response relationship.

 

[Back to Top]

Confidence Interval

The Confidence Interval(CI) is a range of values within which we believe the underlying, true value will be included. Most often, a 95% CI is given, which means that there is a 95% chance the range given includes the true value. One can think of the confidence interval as the range of values representing the estimate of interest, with the calculated estimate being the most probable. If the CI is very wide, the estimate is less reliable. The main factor affecting the width of the CI is the number of people surveyed or otherwise included in the population being measured. So, for small surveys, the CIs are often wide.

When comparing data points such as the answers to survey questions between different age groups or genders, one often looks at the CIs to decide whether or not there are true differences. In general, if the CIs overlap, the numbers are not statistically different.

 

[Back to Top]

Data Variability and Small Numbers

Cancer Rates by County:
The number of new cancer cases reported in a county varies from year to year. Cancer rates, therefore, also vary from year to year. Counties with a small population tend to have a greater degree of variation from year to year. In general, when there are fewer than 30 cases a year, it can be difficult to distinguish between normal variation and meaningful changes in the cancer rates. In this report, multiple years of data are combined when producing the county rates. Although combining years makes the rates more stable, caution must still be used when interpreting county rates. Counties with high rates during one time-period could be low during a different time-period, purely by chance.

Cancer Rate Trends for Specific Sites:
Cancer rate trends for specific cancer sites (body locations) are presented in this report. Cancer rates for less common cancers, such as cervical cancer, are based on a small number of cases. Rates based on a small number of cases tend to be less reliable and should be interpreted with caution.

State at Diagnosis Data:
Stage at diagnosis describes the extent to which the cancer has progressed at the time of diagnosis. Maine staging data are presented from 1995 to 2002. Stage at diagnosis trend data are provided as a way to assess the effects of screening and early detection measures. For cancers that have recommended screening measures, one would hope to see an increase in the percent of local stage diagnoses and a decrease in the percent of regional or distant stage diagnoses over time.

 

[Back to Top]

Defining Disability

Defining disability is a challenge, especially for health data systems that measure the health impact of disabilities. Different data systems classify disabilities differently. Examples include:

  1. The Americans with Disabilities Act of 1990 (ADA) defines a disability as a "physical or mental impairment that substantially limits one or more major life activities".
  2. Social Security Administration, through which services such as Medicaid Health Insurance and financial resources are obtained, defines a disability as a physical or mental impairment that substantially impairs the person's ability to perform work (substantial gainful activity), and the condition must have existed or is expected to continue to exist for at least one year.

Other challenges in defining disabilities include:

  1. Some disabilities, such as mental disorders, may be only temporary, yet others are lifelong.
  2. Often, people in the deaf culture who use American Sign Language do not consider themselves disabled, yet deafness is often included in measurement tools as a disability.

 

[Back to Top]

Disability, How it is Measured

How DISABILITY status is asked in these data sets:

2000 Census
Does this person have any of the following long-lasting conditions:

  • Blindness, deafness, or a severe vision or hearing impairment?
  • A condition that substantially limits one or more basic physical activities such as walking, climbing stairs, reaching, lifting, or carrying?

Because of a physical, mental, or emotional condition lasting six months or more, does this person have any difficulty in doing any of the following activities:

  • Learning, remembering, or concentrating?
  • Dressing, bathing, or getting around inside the home?
  • (Answer if this person is 16 years old or over). Going outside the home alone to shop or visit a doctor's office?
  • (Answer if this person is 16 years old or over). Working at a job or business?

Maine Behavior Risk Factor Surveillance System (BRFSS)
The following question was asked by Maine BRFSS in 2000:

  • During the past 30 days, for about how many days did poor physical or mental health keep you from doing your usual activities, such as self-care, work, or recreation?
    • None, Don't Know / Not Sure, Refused, Number of Days.

The following questions were asked by Maine BRFSS in 2001:

  • Are you limited in any way in any activities because of physical, mental, or emotional problems?
  • During the past 30 days, did poor physical or mental health keep you from doing your usual activities such as self-care, work, or recreation?
  • Do you now have any health problem that requires you to use special equipment, such as a cane, a wheelchair, a special bed, or a special telephone?
    • Yes, No, Don't Know / Not Sure, Refused.

The following questions are going to be asked in Maine BRFSS in 2003:

  • During the past 30 days, for about how many days did poor physical or mental health keep you from doing your usual activities, such as self-care, work, or recreation?
    • Number of days, None, Don't Know / Not Sure, Refused
  • Are you limited in any way in any activities because of physical, mental, or emotional problems?
  • Do you now have any health problem that requires you to use special equipment, such as a cane, a wheelchair, a special bed, or a special telephone?
    • Yes, No, Don't Know / Not Sure, Refused.

A more thorough disabilities module of questions is being planned for Maine BRFSS for a future year.

Treatment Data System (TDS)
Treatment Data Systems for substance abuse treatment

  • Are any special accommodations needed to provide services?
    • Yes or No answer for the following items: hearing, visual, physical, language, or other.

 

[Back to Top]

5 YEAR TRAILING AVERAGE

In some instances, five years of data are combined and the average across the five years presented. This is done to improve the stability of the estimates, which would otherwise fluctuate due to small numbers of occurrences. The five years "trail" or precede the year presented. For example, the 5 year trailing average for 2004 is calculated using data from the years 2000-2004.

 

[Back to Top]

INCIDENCE AND PREVALENCE

In chronic diseases, we commonly measure disease rates with two different methods:

Incidence is the number of newly diagnosed cases of a disease occurring in a population in a given period of time (usually a year).

Prevalence is the total number of cases of a disease in a population at a given point or period in time.

Why is it important to distinguish between these two measures?
Incidence gives us a barometer of how many new cases of a disease are being detected, while prevalence is an indication of both the development of new cases as well as how long people are living with a disease.

For instance, cancer incidence rates are declining across the United States, although not in Maine. Incidence may be declining due to a reduction in causative factors for cancer, such as tobacco addiction. Cancer prevalence rates are rising across the nation, which may be due to improved treatments leading to longer survival.

For short-lived diseases, in which people either die or are cured quickly, incidence and prevalence are very similar. Examples include many acute infectious diseases such as bacterial meningitis or bacterial diarrhea.

 

[Back to Top]

MAKING TWO POPULATIONS COMPARABLE

There are two ways to make two populations comparable when known characteristics are distributed differently between them:

  1. Compare category-specific rates. For example, one can compare cancer mortality rates in 1900 and 2000 for each age group. Age-specific rates for cancer deaths tended to increase only slightly.
  2. Adjust the rates for the characterization; in other words, perform standardization. This can be done by direct and indirect methods, but both methods use a weighted average of category-specific rates. They differ in the source of the weights and rates used. In indirect standardization, rates from a standard population are applied to weights in the study group. In direct standardization, category-specific rates observed are applied to a single standard population. Often the US population for a census year is used as a standard population for comparison.

 

[Back to Top]

MARGIN OF ERROR

The margin of error is a measure of the degree of uncertainty in an estimate, such as prevalence or rate, often due to the estimate stemming from a sampled portion of the population. Consider this, a survey finds that 25 percent of adults in Maine have high blood pressure and the survey's margin of sampling error is plus or minus 1.6 percentage points. The estimate, 25 percent, is considered the most likely value, but we consider a range of plausible values between 23.4 and 26.6 percent (25 - 1.6 and 25 + 1.6). This range is sometimes referred to as the 95 percent confidence interval. In 95 out of 100 samples, we expect the 95 percent confidence interval to include the true value. If the range of estimates based on the margin of error between the state and the district overlap, then it is unlikely that there is a statistically significant difference between the district and the state on that indicator.

 

[Back to Top]

MEDIUM AND MEAN

Median is the 50th percentile, or the middle of the data, the value at which half of the observations are above and half are below.

Mean is the simple average of the data.

When can medians and means be very different from one another when used to describe the same data?
A common example is when there are extreme values, or outliers. For instance, if five people's ages are: 34, 35, 36, 37, and 80, the median is 36 and the mean is 44. Therefore, the advantage of using the median is that it is not affected by extreme values.

 

[Back to Top]

MEASURING QUALITY OF LIFE USING HEALTHY DAYS

The concept of health-related quality of life refers to a person's or group's perceived physical and mental health. Survey data in BRFSS include a set of questions related to health-related quality of life entitled "Healthy Days", which ask a core set of questions related to how individuals have reported feeling in terms of their physical and mental health over the last 30 days. This provides an over-time measure of perception of well-being. Evidence suggests this can be used not only as a measure of individual health but as a proxy for measuring community level health.

Healthy Days measures have been found useful for (1) identifying health disparities in different populations and subgroups and (2) tracking population trends. They can also be used to build broad coalitions around a measure of population health compatible with the World Health Organization's 1948 definition of health: "Health is a state of complete physical, mental, and social well-being, and not merely the absence of disease or infirmity".

 

[Back to Top]

RACE AND ETHNICITY, HOW IT IS ASKED

How RACE and ETHNICITY are asked in these data sets:

2000 Census
RACE:

  • What is the person's race? Mark one or more races:
    White; Black, African American, or Negro; American Indian or Alaskan Indian - print name of enrolled or principal tribe: Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, Native Hawaiian, Guamanian or Chamorro, Samoan, Other Pacific Islander (print race), Other Asian (print race), Some other race (print race).

ETHNICITY:

  • Is this person Spanish / Hispanic / Latino? If yes, check if Mexican, Mexican American, Chicano; Puerto Rican; Cuban; or other Spanish / Hispanic / Latino.
  • What is the person's ancestry or ethnic origin? (print ethnicity; some examples are given, such as Italian, Jamaican, African American, Cambodian, Cape Verdean, Norwegian, etc.)

Behavioral Risk Factor Surveillance System (BRFSS)
RACE:

  • Which one or more of the following would you say is your race?
    • American Indian or Alaskan Native
    • Native Hawaiian or Other Pacific Islander
    • Asian
    • White
    • Black or African American
    • Other  
  • Which one o these groups would you say best represents your race?
    • American, Indian, or Alaskan Native
    • Native, Hawaiian, or Other Pacific Islander
    • Asian
    • White
    • Black or African American
    • Other

ETHNICITY:

  • Are you Hispanic or Latino?
    Yes, No, Don't know / Not sure, Refused

Maine Youth Risk Behavior System (YRBS)

  • How do you describe yourself? (Select one or more responses)
    • American, Indian, or Alaskan Native
    • Native, Hawaiian, or Other Pacific Islander
    • Asian
    • White
    • Black or African American
    • Other
  • Has anyone ever made offensive racial comments or attacked you based on your race or ethnicity - at school or on your way to or from school?
    • yes
    • no

Pregnancy - Related Monitoring System (PRAMS)

  • Information in PRAMS on race and ethnicity is obtained from birth certificates.

Maine Health Data Organization (MHDO)

  • Data is collected from claims forms, which do not ask race and ethnicity.

Cancer Registry (MCR)
RACE:

  • Since hospital records are the source of this information in Maine vary on how and if they ask race, racial status is ascertained in various ways, depending on what information is available from the hospital records.

ETHNICITY:

  • For ethnicity, the possible selections are Mexican, Puerto Rican, Cuban, South or Central American (except Brazil), other Spanish, Spanish Surname, Unknown, and Non-Spanish.

Birth Certificates (BRTH)
RACE:

  • The form asks for a person to specify in blank spaces for race for each parent, giving examples of "American Indian, Black, White, etc"

ETHNICITY:

  • The forms ask to specify in blank spaces for ethnicity for each parent, giving examples of "French, English, Irish, etc."
    Note: If more than one race or ethnicity is entered, only the first is used. However, this practice will change to include all listed races and ethnicity.

Death Certificates (DTH)
RACE:

  • The form asks for person to specify in blank spaces the race, giving examples of "American Indian, Black, White, etc."

ETHNICITY:

  • The form asks to specify in blank spaces the ancestry, giving examples of "French, English, Irish, etc."

Abortion Certificates (AB)
RACE:

  • The form asks for person to check American Indian, Black, White, or Other, and to specify race if "Other".

ETHNICITY:

  • The form asks for the person to fill in Ancestry, with the examples given of French, English, Irish, etc.

Marriage Certificate (MAR)
RACE:

  • The form asks for groom and bride's race to be filled in, giving examples of "American Indian, Black, White, etc."

ETHNICITY:

  • Ethnicity is not asked.

Infectious Disease Reports (ID)
RACE:

  • Race is circled using the following choices: "American Indian or Alaskan Native / Asian or Pacific Islander / Black / White/ Unknown" or race is let blank for the reporting person to fill in.

ETHNICITY:

  • Ethnicity is asked: "Hispanic? Y N" (circle one)

Treatment Data System (TDS)
RACE:

  • White, Black or African American, American Indian or Alaskan Native, Asian, Native Hawaiian or Other Pacific Islander, Other.

ETHNICITY:

  • Not Hispanic or Latino; Hispanic or Latino

Maine Youth Drug and Alcohol Use Survey (MYDAUS)
RACE:

  • What is your race?
    • White, Caucasian or European
    • Other Asian
    • Black or African American
    • Filipino
    • Asian or Pacific Islander
    • Samoan
    • Chinese
    • Hawaiian
    • Chinese
    • Guamanian
    • Korean
    • Other Pacific Islander
    • Asian Indian
    • American Indian or Alaskan Native
    • Cambodian
    • Other (Please specify)
    • Vietnamese

ETHNICITY:

  • Are you Spanish / Hispanic / Latino?
    • No, not Spanish / Hispanic / Latino
    • Yes, Puerto Rican
    • Yes, Mexican American
    • Yes, Cuban
    • Yes, Mexican
    • Yes, Central or South American
    • Yes, Chicano
    • Yes, Other Spanish / Hispanic / Latin

[Back to Top]

RATIO, PROPORTION AND RATE

Ratio is a general term that means there is a numerator (the top number in a fraction) and a denominator (the bottom number in a fraction). Types of ratios include proportions, percentages, and rates.

Proportion is a ratio in which the population in the numerator is also included in the denominator. An example is proportion of women giving birth who have a C-section - 25 out of 100. Proportions are often expressed as a percentage. The above example would be 25%.

Rate
is a ratio in which a measure of time is included in the denominator. An example is the incidence (number of new cases) of breast cancer in a given year.

Ratios, proportions, and rates can easily be confused because they are often used interchangeably, though technically they often should not be. In order to interpret the data correctly, the most important factor is determining exactly what constitutes the numerator and the denominator.

 

[Back to Top]

SENSITIVITY AND SPECIFICITY

Screening tests are evaluated based on their sensitivity and specificity.

  1. A test with high sensitivity means it has a high ability to assure that people who have the disease will test positive and, therefore, will have a high likelihood to avoid missing a true case of the disease.
  2. A test with high specificity means that it has a high ability to assure that a negative test result means people do not have the disease.
  3. Sensitivity and specificity are interrelated. Loosening the criteria that makes a test positive means that more people who have the disease will test positive (increased sensitivity), but so will more people who do not have the disease (decreased specificity, resulting in false positive results).
  4. And, conversely, making more stringent the criteria that makes a test positive means that more people who actually have the disease will test negative and their disease will, therefore, be missed (decreased sensitivity, resulting in false negative results); yet more people who test negative will actually not have the disease (increased specificity).

 

[Back to Top]

SEXUAL ORIENTATION

How SEXUAL ORIENTATION is asked in these data sets:

2000 Census
The 2000 Census does not identify sexual orientation of respondents but does identify some numbers of same sex partners living in the same household. The "householder," the individual in whose name the house is owned or rented, was asked to identify how other people in the household are related to the householder. Categories included spouse, child or other relative of the householder, housemate/roommate, roomer/boarder, and unmarried partner. Those identified as spouse or unmarried partner and hound to be of the same sex were then designated as "same-sex partners". The 2000 Census numbers of same-sex partners are felt to be an undercounting, since some couples would be reluctant to report and since some may not live together.

The 2001 Youth Risk Behavior Surveillance System (YRBSS)
What is your sex?

The person(s) with whom you have had sexual contact during your life is:

  1. I have never had sexual contact
  2. Female
  3. Male
  4. Male and female

Has anyone ever made offensive racial comments or attacked you based on your race or ethnicity - at school or on your way to or from school?

  • yes
  • no

ID reports ask
For Hepatitis, "What is the patient's sexual preference?"
For HIV/AIDS, "What is the gender of sexual partners?"

 

[Back to Top]

MEASURING TOBACCO USE

Tobacco use is measured by two major statewide indicators - the percent of a population that is addicted, commonly known as smoking rates, and tobacco consumption, which is the packs sold per adult.

 

[Back to Top]