BIOSTATISTICS AND EPIDEMIOLOGY

WEEK 1: NATURE OF BIOSTATISTICS  Based on results of prior studies

COURSE INTRODUCTION  Examples of a hypothesis:
 The risk of developing lung cancer
• Epidemiology and biostatistics are the basic sciences
remains constant in the last five
of public health
years
• Public health investigations use quantitative methods,  The use of a cell phone is
which combine the two disciplines of epidemiology and associated with developing brain
biostatistics. tumor
EPIDEMIOLOGY  Conduct a study
 Survey study is used to estimate the
Epidemiology is the study of the distribution and extent of the disease in the population.
determinants of health, disease, or injury in human  Surveillance study is designed to
populations and the application of this study to the monitor or detect specific diseases.
control of health problems.  Observational studies investigate
Examples association between an exposure and a
disease outcome.
 National and local surveillance system (cancer,
 They rely on “natural” allocation of
AIDS,
individuals to exposed or non-
 Occurrence of E. coliO157:H7 outbreak) exposed groups
 Cohort study to investigate the association of  Experimental studies investigate the
cell phone use association between an exposure, often
 And the development of brain tumors therapeutic treatment, and disease
 Survey of individuals who took Cox-2 inhibitors outcome Individuals are “intentionally”
BIOSTATISTICS placed into the treatment groups by the
investigators
Statistics is the science and art of dealing with variation
 Collect data
of data in order to obtain reliable results and
 Numerical facts, measurements, or
conclusions.
observations obtained from an
Biostatistics is the application of statistics to problems investigation to answer a question.
in the biological sciences, health, and medicine  Influences of temporal and seasonal
Examples trends on the reliability and accuracy of
data.
 Computing age-adjusted cancer incidence rates  Examples
to determine trends over time and locality  The number of lung cancer cases
 Calculating statistical measures of the risk of from 1960–2000 in the United
developing brain tumors following cell phone States.
use after adjusting for possible confounding  The number of deaths from
variables cardiovascular diseases in Whites
 Quantifying the relationship between use of and African Americans from 2000–
Cox-2 inhibitors and quality of life 2004.
ROLE OF QUANTITIVE METHODS IN PUBLIC  Describe the observations/data
HEALTH  Descriptive statistical methods provide
an exploratory assessment of the data
 Address a public health question
from a study
 Generate a hypothesis Based on
 Exploratory data analysis techniques
scientific rationale
 Organization and summarization of data
 Based on observations or anecdotal
evidence(not Scientifically tested)  Tables
 Graphs
BIOSTATISTICS AND EPIDEMIOLOGY
 Summary measures Devoted to the summarization and description of data.
 Assess strength of evidence for/against a It includes the construction of graphs, charts, and tables
hypothesis; evaluate the data
 Inferential statistical methods provide Calculation of various descriptive measures such as
a confirmatory data analysis Generalize means/averages, measures of variation, and
conclusions from data from part of a percentiles.
group (sample) to the whole group Describes the important characteristics of a given set of
(population) data
 Assess the strength of the evidence
These statements are examples of DESCRIPTIVE
 Make comparisons
STATISTICS
 Make predictions
 Ask more questions; suggest future  At our health centre, 50 patients were diagnosed
research with angina last year.
 Recommend interventions or preventive
(After Rowntree, 1981)
programs.
 The study results will prove or disprove  Inferential Statistics
the hypothesis, or sometimes fall into a Branch of statistics concerned with using sample data to
grey area of “unsure” make an inference about a population of data
 The study results appear in a peer-
review publication and/or are It includes methods like point estimation, interval
disseminated to the public by other estimation and hypothesis testing which are all based
means on probability theory.
 As a consequence, the policy or action Used to draw conclusions or to make predictions about
can range from developing specific what may happen in other subjects.
regulatory programs to general
Aims to give information about the population by
personal behavioral changes
studying the characteristics of the sample drawn
 Antibiotics reduce the duration of viral throat
infections by 1–2 days.  These statements are examples of INFERENTIAL
 Five per cent of women aged 30– 49 consult STATISTICS
their GP each year with heavy menstrual  Antibiotics reduce the duration of viral throat
bleeding.  infections by 1–2 days. 
 At our health Centre, 50 patients were
 Five per cent of women aged 30–49 consult their
diagnosed with angina last year.
GP each year with heavy menstrual bleeding.
(After Rowntree, 1981) (After Rowntree, 1981)
WHAT ARE STATISTICS? It is important to recognize that when we use statistics
in this way,
Nature of Biostatistics
We never know exactly what the true results in the
 It explores the collection, organization, analysis, population will be. For example,
and interpretation of numerical data. We shall never know how often every woman consults
 It is a branch of science that deals with the her GP (these data are not routinely collected in
identification of the problem, collection, primary care at present), but we can draw a conclusion
presentation, analysis, and interpretation of that is based on a sample of data
data. Example Descriptive and Inferential Statistics
 When the focus is on the biological and health
Consider event of tossing dice.
sciences, we use the term biostatistics
The dice is rolled 100 times and the results are forming
Types of Statistics the sample data.
 Descriptive Statistics
BIOSTATISTICS AND EPIDEMIOLOGY
Descriptive statistics is used to grouping the sample You are interested in “How stress affects mental state
data to the following table: of human beings?”
OUTCOME OF THE ROLL FREQUENCIES IN THE Independent variable -> Stress
SAMPLE DATA Dependent variable -> mental state of human beings
1 10 You can directly manipulate stress levels in your human
2 20 subjects and measures how stress kevels change mental
state.
3 18
Dependent Variable is a variable which is affected or
4 16 influenced by another variable.
5 11 Independent Variable is one which affects or
6 25 influences the dependent variable.

Inferential statistics can now be used to verify whether Example: In a research problem entitled,
the dice is a fair or not. "The Effect of Technology-based Instruction on the
DATA PRESENTATION AND ANALYSIS (Every Students' Mathematics Achievement".
study or experiment yields a set of data)
The independent variable here is the technology-based
Steps in Statistical Investigations instruction, while the dependent variable is the
1. Identification of the problem academic achievement of students.
- We generate the hypothesis Common statistical terms
2. Collection of data
Constant
3. Presentation of data
4. Analysis of data - Quantities that do not vary e.g. in biostatistics,
5. Interpretation of data mean, standard deviation are considered
constant for a population
Collection of data
Variable
 Refers to the different methods and techniques of
gathering data - Characteristics which takes different values for
different person, place or thing such as height,
What is DATA?
weight, blood pressure.
It refers to any information concerning to a population
Categories of Variable
or sample
 Qualitative Variable  Uses categories or
Types of Data (According to Source)
attributes that are distinguished by some
Primary Data nonnumeric characteristics. ( color of the eyes)
 Examples: Interview, registration, experiment,  Quantitative Variable  It is consist of numbers
questionnaire, etc. representing counts or measurements ( height,
weight)
Secondary Data
Types of Quantitative Variable
 Examples: book, journal, newspaper, thesis,
Discrete Data
dissertation, etc.
Types of Data (According to Functional relationship)  Quantitative data which can assume a finite or
countable number of values. (Whole numbers: no. of
Independent Data siblings)
 It refers to any controlling data  Cannot be represented by fractions or decimals
Dependent Data Continuous Data
 It refers to any data that is affected by the controlling  Quantitative data which can assume an infinity or
data many possible values corresponding to a point on a line
EXAMPLE 1: interval
BIOSTATISTICS AND EPIDEMIOLOGY
Examples of conversion of discrete to continuous data
MEASURING DISCRETE CONTINUOUS 3. Interval
Gas Tank Empty/ full Gas volume in It does not have a true-zero value starting point.
tank
There is no absolute zero, unique origin does not exists
Tree Heights Tall/Medium/Short Meters
Interval data are more powerful than ordinal scale due
Performance Poor/Average/Good Pieces/Hour to equality of intervals
Temperature No. of Days > Average Example:
Freezing Temperatures
 Temperature in Fahrenheit
Delivery No of Late Time per
deliveries Delivery  Standardized scores
Scrap No. of pieces > Average part 4. Ratio
Max. Length length
A modified interval level which includes zero as a
starting point.
Measurement of Data Highest level
1. Nominal Relation among Data Types
Data that consist of names, label or categories - Nominal and ordinal -> Nonparametric
commonly used number to categorize data. (qualitative data)
Lowest levels of data - Interval and Ration -> Parametric (quantitative
data)
Examples:
Nonparametric statistics may be used to analyze
Gender, marital status, or any alphabetic/numeric code interval and ratio data measurement.
without intrinsic order or ranking
PRINCIPLES OF SAMPLING
SUBJECT CODE
• Population (N)
Physics P
• Is the set of complete collection or totality of all
Chemistry C possible values of the variable.
Mathematics M • Sample (n)
Biology B • A subset or sub-collection of elements drawn from a
population.

2. Ordinal
Measurement which deal with order and rank.
Second level
Examples
Ranks or grades of students; Quality rating of service or
product
EDUCATION CODE Stages in the selection of a sample

Undergraduate U  Define the population

 Select a sampling frame
Graduate G  Determine if a probability or nonprobability
Post-Graduate P sampling method will be chosen
 Plan procedures for selecting sampling units
Doctorate D  Determine sampling size
 Select actual sampling units
BIOSTATISTICS AND EPIDEMIOLOGY
 Conduct fieldwork Using .05 or .01 as the sampling error, the Sloven’s
Formula may easily be applied
DETERMINING THE SAMPLE
• There is no general rule regarding the sample size. Examples:

• However, the higher the percentage of the sample, Determine the number of samples needed in each:
the higher the validity of the study.
• Sampling techniques
Probability Sampling
 Simple random sampling
• In this technique elements of the sample are selected
through lottery.
 Systematic sampling
• This technique of sampling is done by taking every
element in the population assignment of number as a
part of the sample.
• To select the systematic sample of n elements from a
The bigger population of N element, we divide the N element in
the population, the lesser percentage of the sample is the population in the n groups of kth element.
taken.
• For a specific calculation of the sample for the
purpose of adequate sampling, the use of the Sloven’s
Formula presented below is advised as given by Pagoso  Cluster sampling

Simplified Formula for Proportion • Population under this technique is being divided into
sections (or cluster), randomly select some of these
cluster as the member of the sample size.
 Stratified sampling
• In this technique, the population is subdivided into at
least two different subpopulation (or strata) that share
the same characteristics and then the elements of the
sample are drawn from its stratum proportionately
Example

N = Population;
n = sample
e = margin of error or percentage of error
BIOSTATISTICS AND EPIDEMIOLOGY
Non-Probability Sampling
 Purposive sampling (Judgmental /
Authoritative)
• In this technique, the elements of the sample are
being selected according to the criteria or rules set.
 Quota sampling
• In this technique, the sample size is limited on the
required number or subject in the study
 Convenience sampling (Haphazard /
Opportunity)
• In this technique, the sample are being selected from
a particular place at specified time preferred.

 Snowballing sampling
• In this technique, the researcher asks respondents to
give referrals to other possible respondents.