Public health concepts for interpreting data

1 INCIDENCE AND PREVALENCE

In chronic diseases, we commonly measure disease rates with two different methods:
1.1 Incidence: the number of newly diagnosed cases of a disease occurring in a
population in a given period of time (usually a year).
The incidence rate is the number of persons contracting a disease during a given time
period per 1,000 population at risk. The incidence rate and other morbidity rates vary so
widely that any constant may be used that expresses the rate in a clear manner (from
“per 100” or “percent” to “per 100,000”).

1.2 Prevalence: the total number of cases of a disease in a population at a given point or
period in time.
The prevalence rate is the number of persons who have a particular disease at a given
point in time per 1,000 population. This rate includes all known cases that have not
resulted in death, cure, or remission, as well as new cases developing during the specified
period.
The prevalence rate is a “snapshot” of an existing health situation; it describes the health
status of a population at a point 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 gives
us a barometer of how long people are living with a disease. For instance, cancer
incidence rates are declining across the India, although not in Maine. Incidence may be

Dr Rajan Dubey- dubeyrajan@gmail.com

 Public health concepts for interpreting data
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.

Vital Statistics Measures

Numerator (Events) Denominator
(Pop. at risk)
Birth rate Live Births Midyear pop.
Crude death rate Deaths Midyear pop.
Infant mortality rate Deaths before age 1 Live Births in the year
Age-specific mortality Deaths for a specific age group Pop. in age group
Age-adjusted mortality Crude rate adjusted to a standard pop.

Prevalence and Incidence

Prevalence = Number of Cases in Population /Population

Incidence = Number of New Cases /Total Number at Risk

2. Age and Sex Composition

Age and sex are the most basic characteristics of a population. Every population
has a different age and sex composition—the number and proportion of males and
females in each age group—and this structure can have considerable impact on
the population’s social and economic situation, both present and future.

Developing countries have relatively young populations while most developed

countries have old or “aging” populations. In many developing countries, 40
percent or more of the population is under age 15, while 4 percent is 65 or older.
On the other hand, in all but a few developed countries, less than 25 percent of
the population is under age 15 and more than 10 percent is 65 or older.

Dr Rajan Dubey- dubeyrajan@gmail.com

 Public health concepts for interpreting data
Median Age: - The median age is the age at which exactly half the population is
older and half is younger.

Sex Ratio: - The sex ratio is the ratio of males to females in a given population,
usually expressed as the number of males for every 100 females. After birth, sex
ratios vary because of different patterns of mortality and migration for males and
females within the population.

The Tools of Demography:-

Rate: - The frequency of demographic events in a population during a specified
time period (usually a year) divided by the population “at risk” of the event
occurring during that time period. Rates tell how common it is for a given event to
occur. (For example, in 1997 in Papua New Guinea there were 34 live births per
1,000 population.) Most rates are expressed per 1,000 population. Crude rates are
rates computed for an entire population. Specific rates are computed for a
subgroup, usually the population more nearly approximating the population “at
risk” of the event. (For example, the general fertility rate is the number of births
per 1,000 women ages 15-49.) Thus, rates can be agespecific, sex-specific, race-
specific, occupation-specific, and so on. In practice, some measures that are
referred to as rates would be more accurately termed ratios.

Ratio:- The relation of one population subgroup to the total population or to

another subgroup; that is, one subgroup divided by another. (For example, the sex
ratio in Iran in 1996 was 103 males per 100 females).

3. Fertility
Fertility refers to the number of live births women have. It differs from fecundity,
which refers to the physiological capability of women to reproduce. Fertility is
directly determined by a number of factors that, in turn, are affected by a great
many social, cultural, economic, health, and other environmental factors.

Dr Rajan Dubey- dubeyrajan@gmail.com

 Public health concepts for interpreting data
3.1 Birth Rate-

The birth rate (also called the crude birth rate) indicates the number of live births
per 1,000 populations in a given year.*

*( Most annual rates, such as the birth rate, relate demographic events to the
population at mid-year (July 1), which is considered to be the average population
at risk of the event occurring during the year.)

Births are only one component of population change, and the birth rate should
not be confused with the growth rate, which includes all components of change.

General Fertility Rate-

The general fertility rate (also called the fertility rate) is the number of live births
per 1,000 women ages 15-49 in a given year.*

*(The childbearing ages for women are assumed for statistical purposes to be ages
15- 44 or 15-49)

The general fertility rate is a somewhat more refined measure than the birth rate
because it relates births to the age-sex group at risk of giving birth (usually
defined as women ages 15-49). This refinement helps eliminate distortions that
might arise because of different age and sex distributions among populations.
Thus, the general fertility rate is a better basis to compare fertility levels among
populations than are changes in the crude birth rate.

Total Fertility Rate:-

The total fertility rate (TFR) is the average number of children that would be born
to a woman by the time she ended childbearing if she were to pass through all her
childbearing years conforming to the age-specific fertility rates of a given year.

The TFR sums up, in a single number, the fertility of all women at a given point in
time. In effect, it says: This is the total number of children a woman would have if
the fertility rates for a given year applied to her throughout her reproductive life.
Dr Rajan Dubey- dubeyrajan@gmail.com
 Public health concepts for interpreting data

Child-Woman Ratio:-
The child-woman ratio is the number of children under age 5 per 1,000 women of
childbearing age in a given year. This measure can be calculated from national
censuses or survey data, thereby providing fertility data where birth statistics may
not otherwise be available.

Mortality:- Mortality refers to deaths that occur within a population.

While we all eventually die, the probability of dying during a given time period is
linked to many factors, such as age, sex, race, occupation, and social class. The
incidence of death can reveal much about a population’s standard of living and
health care.

Dr Rajan Dubey- dubeyrajan@gmail.com

 Public health concepts for interpreting data

Death Rate:-The death rate (also called the crude death rate) is the number of
deaths per 1,000 populations in a given year.

Infant Mortality Rate:-

The infant mortality rate is the number of deaths of infants under age 1 per
1,000 live births in a given year.

Dr Rajan Dubey- dubeyrajan@gmail.com

 Public health concepts for interpreting data

Share of Under-Five Mortality in India

Neonatal deaths Infant deaths within 7 days of birth 38%

50%
Infant deaths
Infant deaths between 75%
7 days of birth and within 29 days 12%

Infant deaths between 29 days 25%

and within one year of birth

Child deaths between one 25%

year and within five years of birth

CAUSES OF NEONATAL DEATHS

About 3/4th of
the neonates
dying are Low
Birth Weight

ICMR 2006
Some aspects of Neonatal deaths -
• 10-15% of the babies are preterm
• 25-30% are low birth weight (< 2500 grams)

Dr Rajan Dubey- dubeyrajan@gmail.com

 Public health concepts for interpreting data
Maternal Mortality Ratio:-

The maternal mortality ratio is the number of women who die as a result of
complications of pregnancy or childbearing in a given year per 100,000 live births
in that year. Deaths due to complications of spontaneous or induced abortions are
included.

This measure is sometimes referred to as the maternal mortality rate; it is best to

specify the denominator when using either measure. A true maternal mortality
rate would divide the number of maternal deaths by the number of women of
childbearing age in the population.
In practice, a maternal death is defined as the death of a woman while pregnant or
within 42 days of termination of pregnancy from any cause related to or
aggravated by the pregnancy or its management but not from accidental or
incidental causes.

Medical Causes of Death of Women:-

ot
Classifiable Abortion
9% 9%
Toxemia
Perperal 8%
sepsis
16%

Anemia
Obstructed 19%
labour
10%

Dr Rajan Dubey- dubeyrajan@gmail.com

Haemorrhage
29%
 Public health concepts for interpreting data
The Three Delays
Delay in deciding to seek Delay in reaching Delay in providing care by the
care appropriate care facilities

Delay by families in Delay in reaching the Delays at in getting prompt

recognizing Danger Signs
and Deciding to Seek Care
due to:
• Inadequate woman's
knowledge of danger signs
• women’s status in the
household
• Inability to pay
facility:
Due to distances from
village to facility
• Poor road conditions
• on availability of public
transport and
communication
care at the facility:
• on-availability of
providers
• on-availability of critical
supplies like blood and
drugs
• Indifferent attitude of the
providers to the clients

• Strengthen skills of
ASHAs to counsel mother
to identify danger signs
• Develop a supervision
plan for better
communication with the
families
• Ensure that ASHA referral
fund is used appropriately
• Ensure that ASHA
leverages other funds for
referral
• Plan with the MoIC for
better preparedness
• Develop a supervision
plan for Community level
monitoring

Timing of Maternal Death after delivery

10.3 44.2
23.0

Within 24 hours
1 day to 7 days
8 days to 42 day
22.5
Don't know

Dr Rajan Dubey- dubeyrajan@gmail.com

 Public health concepts for interpreting data
{Understanding Reports and Research Data for
Health}
WHY ASSOCIATION DOES NOT NECESSARILY MEAN CAUSATION:
Epidemiology often looks 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 mean causation.
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 sets of data) and a chi-square
test (which also compares the outcomes of two sets of data). 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 assured 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

Dr Rajan Dubey- dubeyrajan@gmail.com

 Public health concepts for interpreting data
comparing chronic disease rates between different time frames or geographical
areas, the rates should be age-adjusted in order to make them comparable.
Alternatively, disease rates only for pertinent age groups should be compared.
4. 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.

Relative Risk is a common measure used to show the magnitude of an association, and is
often examined in making a judgment pertaining to causality. Relative risk can be
expressed in different ways, depending on the study design. For instance, in a Cohort
study (in which participants are selected based on their exposure or non-exposure to a
possible risk factor for a disease), relative risk is expressed as the incidence of a disease
in those exposed to a possible risk factor divided by the incidence of the same disease in
those not exposed to the risk factor.
In a Case Control study, (in which the participants are selected based on their disease
status), the relative risk can be expressed as an Odds Ratio. The Odds Ratio is the ratio of
the odds of exposure among the cases to that among the controls.
Hennekens, C., J. Buring, Epidemiology in Medicine, Little, Brown, and Company, 1987.

SENSITIVITY AND SPECIFICITY

• Screening tests are evaluated based on their sensitivity and specificity.
• 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.
• 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.
• 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).

Dr Rajan Dubey- dubeyrajan@gmail.com

 Public health concepts for interpreting data
• 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).

MEDIAN AND MEANS

1. 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.
2. Mean is the average of the data.
3. 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. However, this can also be a disadvantage because it
provides no information about distribution of the values since its derivation is
based on rank.

WHY CAN’T WE SIMPLY COMPARE CRUDE DISEASE RATES?

One common problem comparing disease or death rates between populations is that the
groups may differ with respect to characteristics such as age, sex, or race that may affect
the overall rate of disease.
These differences can make crude rates not comparable. For instance, crude death rates
due to cancer in the India have dramatically increased over the past 100 years. However,
the population has also aged. Since cancer mortality rates rise dramatically with age, the
higher crude death rates seen now are at least in part due to overall aging of the
population.

Dr Rajan Dubey- dubeyrajan@gmail.com

 Public health concepts for interpreting data
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 India population for a census year is used as a standard
population for comparison.

CONFIDENCE INTERVALS

The Confidence Interval (CI) is a range of values that represents the true value of a
statistic. Most often, a 95% CI is given, which means that there is a 95% chance
the range given includes the true value. 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. One common method for dealing with wide and overlapping
CIs is to compile multiple years of data together to create a sufficiently large sample
size.

Dr Rajan Dubey- dubeyrajan@gmail.com