Introduction to Applied Biostatistics in public health

An overview introduction to applied statistics

1. Introduction to Public Health (COBERS).
Course Code: CHS 3303 :

Biostatistics concepts, source of data and Scales of measurements and variable
classification


Fredrick Makumbi, PhD
School of Public Health
Makerere University
December 22
nd
2020

2. https://online.regiscollege.edu/blog/what-is-biostatistics-definition-and-application-of-a-key-medical-term"%
“… genomic tests that make possible
individualized treatment of cancerous tumors,
whole-genome sequencing to predict a
patient’s likelihood of contracting specific
diseases, and the use of DNA tests in
pharmacogenomics to identify the medications
that are most compatible with a patient’s
genetic makeup”.
More Efficient Health Care Delivery
“The rigorous and objective
conversion of medical and/or
biological observations into
knowledge.”
Individualized Care’s Potential to Improve
Patient Outcomes
Applying Biostatistics to Improve the Quality of Health Care

3. Applying Biostatistics to Improve the Quality of Health Care
The National Cancer Institute defines biostatistics as “the science of collecting and
analyzing biologic or health data using statistical methods.”
Epidemiologic studies
https://www.kolabtree.com/blog/six-ways-biostatistics-transforming-healthcare/
DRUGS
https://www.kolabtree.com/blog/six-ways-biostatistics-transforming-healthcare/
World Health Statistics 2010

5. Course Objectives
1 To define the concepts & common terminologies in
biostatistics
2 To explain sources of data and application
biostatistics to health sciences
3 To describe the different types of
a) Variables
b) Data
c) Scales of measurements

6. WHAT YOU WILL LEARN
1) Calculate summary statistics from public health and
biomedical data
2) Interpret written and visual presentations of statistical
data
3) Evaluate and interpret results of various regression
methods
4) Choose the most appropriate statistical method to
answer your research question

7. What is Biostatistics?

May be defined as;
the application of the mathematical tools used in statistics to
the fields of biological sciences and medicine
the application of statistical principles to questions and
problems in medicine, public health or biology,
environmental& laboratory sciences.
Translating data into meaningful information that can then be
used to make logical decisions
– Conceptualization, collection, analysis, and interpretation and
presentation of data in meaningful way
7

8. Translating data into meaningful information
that can then be used to make logical decisions
Biostatistics: Evidence based/
supported decision making

9. Example: How many patients do I put on staff at
any given time period?
i) Too many workers; unnecessary labor costs may go up
ii) Too few workers; poor customer service outcomes –may be fatal for
patients
https://www.datapine.com/blog/big-data-examples-in-healthcare/

10. Example, USA Overdoses:
Overdoses from misused opioids have caused more
accidental deaths than road accidents
https://www.datapine.com/blog/big-data-examples-in-healthcare/

11. Example: As a hospital director/facility manager
Data on every facility division, the attendance, the costs incurred,
will be of a great help to run the facility smoothly.
https://www.datapine.com/blog/big-data-examples-in-healthcare/

12. Role of Biostatistician
E To guide the design of studies (experiment/
surveys) prior to data collection

: To analyze data using appropriate statistical
procedures and techniques

* To present and interpret the results to researchers
and other decision makers including journals,
workshops or any meetings

13. Concerns of Biostatistics
• To conceptualize, collect, organize, describe
and summarise information/data
• To make inferences from the observations/
sample data given uncertainty
• To identify associations, relationships, or
differences between sets of observations
13

14. Concerns of Biostatistics

15. Concerns of Biostatistics

16. Concerns of Biostatistics

17. Sources of
data
Routine
records (e.g. HMIS)
Surveys (e.g.DHS)
Community/Clinical
Experiments
(e.g. Circumcision trials)
Comprehensive
(e.g.DHS-national surveys)
Sample
(Cohort of patients in care)
Sources of Data on Health

18. Sources of Data on Health
https://www.google.com/search?
sa=N&q=different+sources+of+data+on+health+and+disease&tbm=isch&source=univ&ved=2ahUKEwirgqmdnM_iAhUFy4UKHdcgCW44ChCwBHoECAAQAQ
&biw=1440&bih=746#imgrc=rEBd2kIMJQWXkM:

19. Sources of Data on Health
http://www.aho.afro.who.int/profiles_information/index.php/TopicView:Data_sources_and_generation

20. Sources of Data on Health
http://collagenrestores.com/activity-diagram-of-hospital-management-system-image/information-systems-in-health-care-health-care-service-delivery/

21. Concepts and common terminologies
Variable
§ A quantity that may vary from object to object
§ Any entity that may take on different values
across individuals and time
§ Or simply “what is being observed or measured ”
e.g. birthweight, Malaria infection, which varies from
person to person.
Value
§ Contents of a variable (quantitative or
qualitative)
Birthweight(in kg)=1.36, 3. 59, 2.82, 1.53

22. Concepts and common terminologies
Population:
a set which includes all
measurements of interest
to the researcher
(The collection of all responses,
measurements, or counts that are
of interest)

Sample:
A subset of the population
Population
sample

23. Concepts and common terminologies
Sampling Design and Procedures 7 th Session of Marketing Reseach ...
‪SlidePlayer
2 [email protected] Population Sample Element ...‬
https://slideplayer.com/slide/8603133/

24. Concepts and common terminologies
https://nursekey.com/sampling-2/

25. Concepts and common terminologies
Sample (or data set)
§ A collection of values of one or more variables as
part of a larger population (a part of the population
from which we actually collect information)

Element (record/observation)
§ A member of the sample
Sample space or population
§ A set of all possible values of a variable (the entire
group from which we want information).

26. Concepts and common terminologies
Statistical inference
§ “the attempt to draw conclusions concerning all
members (population) from observations of only
some of them (sample)” (Rune 1959)
A parameter
§ A numerical descriptor of a population e.g.
population mean (can be age, height, weight, etc.).

A statistic
§ A numerical descriptor/characteristic of a sample
e.g. sample mean.

27. Statistical inference
(Sample to population)

28. What informs type of analysis, and
data presentation

29. 1 Types of variables
i) Qualitative (Categorical) Variables
ii) Quantitative variables

30. i. Qualitative (Categorical) Variables
§ Has values that are intrinsically non-
numerical (can only be put into
categorical)
e.g. color of skin, gender, tribe, country
of residence, model of vehicle you use

31. ii. Quantitative variables

§ Has values that are intrinsically numerical
e.g. birth weight (in kg)=2.3, 3.8, 1.8, 4.2;
height, age

32. 2) Type of Quantitative data
i) Discrete data
ii) Continuous data

33. i) Discrete data
• Takes on only certain numeric values such as
count of events, number of children (usually
whole numbers)
• Have a finite number of possible values, with
space between values on a number line
• E.g. Number of teachers, ANC, emergencies
registered

34. ii) Continuous data
• Takes on values including decimal places,
within a range (SBP 90mmHg, BMI 26.1kg/m
2
)
• Values fall on a continuum, may have
fractions/ decimals

35. ii) Continuous data
Measuring weight
• continuous

36. ii) Continuous data

37. ii) Continuous data

38. 3 Scale of measurement
i) Nominal
ii) Ordinal
iii) Interval
iv) Ratio

39. Scale of measurement

40. i. Nominal/Classificatory Scale

1) Observations fall into categories
2) Named categories (labels) with no implied order
3)  (e.g. cars by model, profession, marital status, etc…)
4) Lowest form of measurement

41. i. Nominal/Classificatory Scale

• There is really no ‘scale’ because it does not scale objects along any
dimension
• However, simply labels objects
• Gender is nominal scale: Male = 1&Female = 2
• A new label Male=3 & Female=1 does not change meaning of original
variable gender

42. ii. Ordinal/Ranking Scale
1) Ordered categories, but differences between
categories can not be considered to be equal
2) Ordered data, lowest to highest (e.g. positions,
scores such as Agree-disagree, Military ranks etc…)

43. iii. Interval
• An interval scale is a scale on which equal intervals
between objects, represent equal differences
• Interval relationships are meaningful
• But, we can’t defend ratio relationships,
• because zero point is arbitrary
• E.g. IQ, Temperature (a zero difference in degree
centigrade is not a 0 in Fahrenheit

44. iii. Interval
A 10-degree difference has the same meaning
anywhere along the Fahrenheit scale
E.g. the difference between 10 and 20 degrees
is the same as between 80 and 90 degrees in
Fahrenheit
But, we can’t say that 80 degrees is twice as
hot as 40 degrees

45. iii. Interval
• 0
0
C is Not same as 0
0
F
• No ‘true’ zero,
only an ‘arbitrary’
zero

46. iv. Ratio
• Equal intervals between values and a meaningful zero point
(height, weight, BP).
•  A zero difference is same irrespective of unit of measurement
(e.g. height of 0 cm, is same as 0 meters)
• Ratios are meaningful
• 20 kgs are twice heavy as 10 kgs
• Difference between 10kgs of beef and 10kgs of fish (zero) is
same as difference in corresponding weight of meat and fish in
grams

47. iv. Ratio
Measuring weight
• Ratio scale

48. How data are analyzed and presented depend
on type of data and scale of measurement

49. SUMMARIZING DATA
Measures of central tendency/location
Mean
Median
Mode
Measures of dispersion/Variability
Range
Percentile
Standard deviation
Interquartile range
Graphical display of data
Box plot
Stem and leaf

50. Graphics for continuous data
Histogram
Box plot
Stem and leaf
Line graphs
Infographics
Scatter plots

51. Line graphs (continuous data-trends)
Figure 3 Changes in systolic blood pressure (SBP)

52. Graphics for Discrete data
Bar
Pie
Infographics

53. Contribution of LARC to mCPR in North
32.4% of modern method
was due to Long-term
43.7% of modern method
was due to Long-term
Stack Bar graph

54. Acknowledgement

55. Reference Books
Wayne W. Daniel - Biostatistics: A Foundation For Analysis In The Health
Sciences Books:
Statistical Methods in Medical Research (P. Armitage, G.Berry &J.N.S Mathews)