BIOSTATISTICS &

DEMOGRAPHY
(CMH1111)

Lecturer: Ms Fatumah Nakku

 Announcements
• Hours/days
– Friday (4 to 6 pm)
– Saturday (11 to 1 pm)
• Mode of delivery
– Lectures
– Assignments
• Mode of Assessment
– Progressive assessment [40%] includes mid-
semester exam, assignments and attendance;
– Final written examination [60%]
 COURSE OBJECTIVES

This course introduces students to the basic

statistical methods of data collection, analysis,
and interpretation.

The aim of the Biostatistics and Demography

course is to develop skills in summarization,
interpretation and presentation of
medical/public health data.
 References/Resources
• Kirkwood, RB and Sterne, JAC 2003. Essentials of
Medical Statistics, 2nd Edition. Blackwell Science,
Massachusetts
• Martin Bland 2000. An Introduction to Medical
Statistics, Third Edition. Oxford University Press, Great
Britain
• Poston Jr., DL and Bouvier, LF 2010. Population and
Society: An Introduction to Demography, First Edition.
Cambridge University Press, New York.
• Other Medical Statistics Books in the library
•  Internet resources
 Lecture 1: Outline
o DEFINITIONS – Biostatistics, Data, Population, Sample and Variables

o TYPES OF STATISTICS- Descriptive and Inferential statistics

o RELEVANCE OF BIOSTATISTICS

o USES OF DATA

o TYPES OF DATA

o SOURCES OF DATA
 Definitions: Biostatistics

Biostatistics is the theory and techniques for describing,

analyzing and interpreting health-related data, and using the data
to estimate the strength of associations between variables.
 
Biostatistics is concerned with the collection of data, the
analysis of data and presentation of the summarized data,
interpretation and making conclusions based on the analysis.
 
 Definitions: Data
Data are numbers; facts obtained from measurement, counting,
questioning/interviewing or observations. Data is evidence of
occurrence of events or existence of phenomena.

 Data is a collection of facts /documented information or

evidence of any kind from which conclusions may be
drawn.
 Data is factual information organized for analysis or used
to make decisions. (thus the terms ‘data’ and ‘statistics’ are
sometimes used interchangeably)
 Examples of health-related data is information describing
patient characteristics which is recorded in patient charts
or registers in a Clinic, Health Centre or Hospital.
 Definitions: Variable

A variable is any measurable characteristic or attribute that

takes on different data values in different persons, places or
things.

Examples of variables include sex, height, heart rate, blood

pressure.

Data values may vary;

- from time to time for the same individual,
- from one individual to another,
- or from one place to another.
 Definitions: Population
A population is the totality of all elements with a defined
characteristic.

Populations may be identified by geographical location, by time, or

by other characteristics of interest such as gender. A combination
of all the three pieces of information is often used to identify a
given population.
 
Examples of populations
-All the people living in Jinja Municipality in 2019.
 Definitions: Sample

A sample is a subset of a population.

Often, data is collected on samples of people,

animals, products, and not whole populations.
 TYPES OF STATISTICS
Descriptive statistics- A set of methods used to make
descriptions of the sample data that we have collected
(summarize the data, identify patterns and relationships in the
data).

 
Inferential statistics- A set of methods used to make
estimates of population characteristics, assess for statistical
significance of relationships between variables, or make
predictions.
 RELEVANCE OF
BIOSTATISTICS
 There is a natural variation in biological and non-biological
processes. There is a variability in biological characteristics,
socio-economic and demographic attributes, perceptions etc
related to health and health behavior.
 How can we describe the characteristics or attributes of a
group of people if they vary from one person to another, or
from place to place?
 How ca we describe or explain the differences between groups
of people if their characteristics and their experiences of
health-related events vary?

 Statistical methods are applied to summarize data with inherent

variability, to assess significance of relationships, and to make
generalizations about the population.
Variability of data- Examples
Person ID Age Haemoglobin Sex
A001 25 13.7 M
A002 19 12.9 F
A003 22 10.5 F
A004 20 12.9 M
A005 24 13.5 M
A006 21 12.1 M
A007 35 15.1 F
A008 21 15.1 F
A009 23 11.4 F
A010 26 8.8 M
 USES OF DATA/STATISTICS
 The demand for quantitative data to support decision-making is
increasing. Statistics or summarized data may be used to:

Know the status of population health and health service

delivery; to identify what needs to be improved.

Plan and administer resources - purchase of hospital supplies,

deployment of teachers/doctors/nurses
 USES OF DATA/STATISTICS (ctd…)

Statistics or data may be used to:

Set financing priorities- to determine resource allocation to

programs and administrative units

Set targets and benchmarks for the future

Monitor performance of health programs

Improve quality of services/products etc

 SOURCES OF DATA

Sources of data are classified into two major categories:

1- Conducting observational or experimental studies (new data)

2-Existing databases created during routine service delivery

(including registers), databases of previous studies or existing
reports (secondary data).
Determining the appropriate data
source
 The choice of the source of data depends on the:

 Problem under investigation

 Study question /research question

 Availability of resources (time, personnel, cost of other

resources)

 Required level of data quality

 Existing data sources
Already existing data were collected for some other purpose
other than addressing the issue or research problem currently
under study. Specific examples of existing data sources include:

– Family planning service records in health facilities

– Police records of criminal offenses, motor accidents etc

– Records of travellers kept by the immigration department

– Publications such as the Uganda Demographic and Health

Survey reports
 Limitations of using existing data
 Missing variables of interest. For example, if the researcher wants to
determine the percentage of women family planning users who agree with
their spouses to use modern contraceptives, the family planning register at
the health facility may not be helpful because information on couple-
decisions is missing.
 
 Incompleteness of data- for example patients’ age is not recorded for all
patients treated in a hospital
 
 Inaccuracy due to reporting and recording errors.

 Lack of representativeness- People whose seek services may not be

representative of all those affected by the problem of interest. They have
been self-selected by other factors: for example, severity of disease, level of
education etc. This lack of representativeness affects generalizability of
findings.
 Collecting New Data
 New data is gathered for the first time to answer a specific
research question or to solve a problem at hand.
Approaches for collecting new data include:

-Conducting surveys and census (Observational studies)

-Conducting experimental studies such as clinical trials of

new drugs, agricultural experiments on new varieties of
animals/crops/fertilizers etc
 Limitations of collecting new data
Although new data collected for a specific purpose is expected to
be the most reliable and accurate, there are various limitations.

Data collection is expensive and time consuming.

Sometimes samples studied are not-representative of the target

population which limits generalizability of findings.

It may be impossible to collect new data due to ethical concerns

and regulations.
 Types of Data
Data types

Qualitative/Categorical Quantitative

Nominal Ordinal Continuous

Discrete
(Measured on a continuum)
(count data)

Ratio Interval
 Types of Data
Qualitative/Categorical variables
 
Nominal - Possible values are mutually exclusive qualitative
descriptions that are not ordered

e.g. Marital status [single, married, widowed, separated]

Ordinal data are ordered but the difference between values is

not meaningful.
e.g. Socio-economic status [low, medium, high]
Categorical Data- Data management & Analysis issues

Nominal data not ordered; Ordinal data has natural order. In

practice, for a given categorical variable, the possible values are
assigned numerical codes. E.g.

Marital status: Single(1), Married (2), Widowed (3), Separated (4)

Socio-economic status: Low SES (1), Medium SES (2), High SES (3)
Subjective rating (e.g Quality of health services): Terrible (0), Poor (1), Fair (2),
Good (3), Excellent (5)

However such numbers (codes) do not have a numerical value in

regard to magnitude. The mathematical difference between
numerical codes or categories is not meaningful.

e.g. SES: Medium(2)-Low(1)=1, High(3)-Medium(2)=1, however the economic

gap between low and Medium SES is not the same as the economic gap between
Medium and High SES.
 Types of Data
Quantitative variables
 
Discrete – data arise out of a counting process eg number of
children seen at the clinic per day

Continuous- data arise out of a measurement process eg weight,

height. The precision of the measurements recorded are only
limited by the precision of the measuring instrument.
 Types of Data
Continuous variables/data are further classified as interval or
ratio.
 
Interval data -e.g temperature, PH, IQ score. The difference
between two values is meaningful e.g. the difference between 36oC
and 37oC and the difference between 38oC and 39oC. A change in
temperature by 1oC represent the same amount of heat. However
the ratio between two values is not meaningful.

Ratio data- both the difference and ratio between two values are
meaningful. A ratio variable has a clear zero point. e.g. weight,
distance. Some objects can be weightless but they cannot have
negative weight.
 Databases and Data- Examples

IN A DATABASE, DATA IS ORGANISED IN ROWS AND COLUMNS

Variables for the Obesity Study

This study was conducted to assess the prevalence of overweight
among school-going children in Jinja Municipality.

Championsaleague2015.xlsx(Database of games played and scores

during the Champions league in 2015.

Development indicators for Uganda (From World Bank Databases)

 Some of the Variables for the Obesity Study

Identify the different types of variables you can see in the data set.
Championsaleague2015.xlsx
Development indicators for Uganda
 Data collection methods
Data Collection Methods Data Collection Tools
Interviews
Interviewer administered Face- Questionnaire (A set of structured
to-face, Phone interviews or semi-structured questions with
Self-administered: pre-selected possible answers)
Posted mail, internet-based
  Checklist (List of items to observe)
Observation
Physical measurement
Case report form
Document review Checklist, Data form
 Guidelines for developing research questions
FINER Criteria
A good research question is:-
 
Feasible (there is adequate number of subjects, adequate technical
expertise, it is affordable in time and money, and is manageable in
scope)
Interesting to the investigator, other scientists and community
Novel (provides new findings, confirms or refutes previous
findings, extends previous findings)

Ethical (meets ethical standards and regulations)

Relevant (to growth of scientific knowledge, to clinical practice
and health policy, to future research directions)
Guidelines for developing research questions
PICOT Criteria
Population (patients)- What specific (patient) population are you
interested in? The sample of subjects you wish to recruit for your
study?
Intervention (for intervention studies only)- What is your
investigational intervention (product)?
Comparison group- What is the main alternative to compare
with the investigational intervention?
Outcome of interest- What do you intent to accomplish, to
measure, improve or affect?
Time- What is the appropriate follow-up time to assess the
outcome or the duration of data collection?
 FORMULATING RESEARCH QUESTIONS

EXERCISE (BRAINSTORMING)

WORKING IN GROUPS OF 2-4 STUDENTS:

-IDENTIFY A PROBLEM YOU HAVE OBSERVED OR WHICH

SHOULD BE SOLVED, AND FORMULATE A RESEARCH QUESTION
USING THE GUIDELINES PROVIDED?

-IDENTIFY KEY VARIABLES FOR YOUR RESEARCH

-IDENTIFY DATA SOURCES

-CHOOSE THE APPROPRIATE DATA COLLECTION METHOD(S)

-END-