2. RESEARCH
1. Research is a scientific and systematic search
for pertinent information on a specific topic
2. It is a careful investigation or enquiry
especially in search of new fact
3. It is a search for knowledge
4. It implies defining, redefining problems,
formulating hypothesis, collecting, organizing
and evaluating data
5. It is systematic approach concerning
generalization and formulation of a theory
6. A careful investigation/enquiry specially
through search for new facts in any branch of
knowledge
3. Type of research
Descriptive vs. Analytical: Descriptive research
includes surveys and it is the description of the
status whereas analytical research is analysing facts
and information to make critical evaluation of the
material
Applied vs. Fundamental:. Applied research aims
at finding a solution for an immediate problem
and fundamental research is concerned with
formulation/generalisation of a theory
Qualitative and Quantitative: Quantitative
research is based on measurement of quantity
whereas qualitative research is concerned with
on qualitative phenomena.
4. RESEARCH DESIGN
It is a comprehensive master plan of the research
study. It is the plan for collecting and utilising
data so that the desired information is
obtained
The task of defining the research problem is the
preparation of the research project popularly
known as research design
Decision regarding what, where, when, how
much, by what means concerning a research
study constitute a research design
7. Randomised (Controlled) Clinical Trial
A clinical trial is a planned experiment designed to
assess the efficacy of a treatment in humans by
comparing the outcomes in a group of patients
treated with a test treatment with those observed in
a comparable group of patients receiving a control
treatment where patients in both groups are
enrolled, treated and followed over the same period.
Curtis L Meinert: Clinical Trials, Oxford Univ Press, 1986
8. Experimental versus Quasi-Experimental:
Experimental studies: measure the impact
of treatments against a comparison or control
group. Comparability is more precisely
established in experimental research, through
random assignment. Experimental studies are for
proving the casual relationship.
In contrast, quasi-experimental research design
is often employed in field settings where people
or groups cannot be randomly assigned
for either ethical or practical reasons
9. Design decisions
• What is the study? Identify topic of interest and
formulate of hypotheses about the answer
• Why/where is the study conducted?
• What is the type of data?
• Where is the data collected from?
• What is the time period?
• What is the sample size and design? (how many
samples are required for a significant effect)
• What is the method of data collection?
• How is the data analysed?
• How is the data reported?
10. Definition of a bias
Systematic error in the protocol of a study that
leads to a distortion of measurement
affecting internal validity
The concept of bias is the lack of internal validity
or incorrect assessment. Bias is said to occur
when the results of the study deviate from the
truth.
Selection bias and information bias
11. Types of biases
• Selection biases
▪ Biases in the way subjects enter a study
• Information biases
▪ Biases in the way information is collected after
inclusion in a study
Biase
s
12. • To remove the potential bias in the
allocation of participants in the intervention
and control groups
• The groups are comparable
• Amenable for statistical analyses
(Randomisation or random allocation is the
process in which the subjects/participants
are allocated to the experimental and
control groups by random assignment)
Why do we randomise?
13. Reliability and validity
Reliability: refers to the ability of the
measurement to give the same results (or the
extent to which the results agree) with
repeated measurement of the same setting
(precision)
Validity: refers to the ability of a measurement
to be correct on the average. How close it is to
the real value (accuracy)
14. • Independent variables that are not related to
the purpose of the study, but may affect the
dependent variable.
• Effect of confounding variables can be avoided
by matching.
E.g. To find out whether teenage mothers are at a
higher risk of having low-birth weight infants.
Mother’s nutritional habits and prenatal care
will be the confounding factors, the effect of
which can be avoided by matching the two
groups, young and old mothers, for these
factors.
Extraneous variables
15. Key areas
• Before and After without control design
• After only with Control design
• Before and after with control design
16. Before and After without control design
• The one group pretest – post test
▪ A pre-experimental design
▪ Uses single test group
▪ Dependent variable is measured before and after
treatment
▪ The effect of treatment is equal to difference in
level of phenomenon
Treatment
Treatment effect = Y-X
X Y
17. After only with control design
The pretest – post test control group
Control group design – A true experimental design.
Uses two groups with Randomization.
The phenomenon is measured only After
Treatment introduced to experimental group.
Dependent variable is measured in both the groups
18. After only with control design
Experimental group X treatment Y
Control group X No treatment Z
Treatment effect = Y-Z
Assumes that two groups are identical and
phenomenon is measured only after.
19. Before – and – after with control design
• Non-equivalent control group design
(a quasi experimental design)
• Matching of elements in the group
• Uses two groups
• Dependent variable is measured in both the
groups for an identical period
• Treatment introduced in experimental/ test
group area
• Dependent variable is again measured in
both for an identical period
20. Before – and – after with control design
Time period1 Time period 2
Experimental
Group (Treatment)
Control
Group (No treatment)
Therefore treatment effect = (Y-X) – (Z-A)
Using more than two groups, we can vary
treatment intensities on experimental
groups.
X Y
Z
A
21. Example1: Erythrocyte sedimentation rate [ESR] was
measured before and after treatment of infection in
10 patients. We wish to examine whether the
treatment conferred any significant benefit.
Patient
No.
ESR – 1 hour (mm)
Before treatment After treatment
1 25 8
2 43 10
3 38 6
4 20 7
5 41 10
6 48 5
7 15 8
8 28 9
9 35 4
10 33 3
Total 326 70
22. Example-2
A trial was conducted to evaluate the efficacy of a
new drug at four different centres employing the same
protocol. The cure rate obtained with the new and
standard drugs at the four centres are given below
Centre Cure rate
New drug Standard drug
A 68%(40/59) 58%(36/62)
B 67%(47/70) 59%(44/74)
C 69%(45/65) 62%(43/69)
D 79%(55/70) 72%(48/67)
Total 71%(187/264) 63%(171/272)
23. Example-3
A nutrition intervention survey was planned. 20 villages
were randomly selected and a base line survey of
nutritional status carried out. A nutrtional supplement
programme was introduced in 10 of the villages and in the
other ten there was no intervention. After 1 year again a
followup survey was carried out to evaluate the effect of
nutritional supplementation programme
25. Observational and interventional studies
• Observational studies
(e.g., What is the incidence of measles?)
(e.g., What are the risk factors for TB?)
• Interventional studies
(e.g., What is the effect of an intervention?)
26. Cohort studies
Definition: A cohort can be described a group
of subjects which is precisely defined at the
outset and of which the composition remains
unchanged.
Cohort studies Involve following group of
subjects over a period of time. Subjects are
defined on the basis of presence or absence
of the exposure to suspected risk factor for
a disease. The cohort is free from disease.
The cohort is followed over a period of time
to assess the outcome of interest.
28. Type of cohort studies
• Prospective: In a prospective cohort study
the investigator chooses or defines a sample
of subjects who do not have the outcome
• Retrospective: In a retrospective study the
patients are enrolled on the basis of their
exposure but the outcome have all
happened in the past
29. Advantage of retrospective study
• Less costly and quick
• Effective for disease with long latent period
• Availability of records- a limitation
• Potential factors (confounding) can be
assessed
30. Retrospective cohorts studies
• Recruitment of study participants
• Retrospective assessment
▪ Collection of information about exposure
▪ Collection of information about illness
31. Basic relation between exposure,
time and outcome
Exposure
Outcomes
(e.g., Disease)
Time
exposure
period
(Time during which
exposure occurs)
Time at risk for
exposure effects
Understand that dynamic when designing the cohort
32. ill Non ill Total
Exposed a b a+b
Non exposed c d c+d
Total a+c b+d a+b+c+d
Relative Risk (RR) is measures the association
between exposure and risk of certain outcome
RR=inc. among exposed / inc. among non-exposed
=a/(a+b) / c/(c+d)
Presentation of the data of an analytical
cohort study in a 2 x 2 table
33. ill Non ill Total
Exposed a b a+b
Non exposed c d c+d
Total a+c b+d a+b+c+d
Relative Risk (RR)=incidence among exposed /
incidence among non exposed=a/(a+b) / c/(c+d)
Relative Risk (RR) is a measure of the
strength of association between an outcome
and exposure to a risk factor and points
towards causation and useful in research for
the etiology of a disease
34. Take home messages
• Cohorts bring together persons sharing a
common experience to follow them over
time
• The cohorts may be prospective or
retrospective
• Cohorts allow precise assessment of
exposure over time
• Cohorts allow experimental designs
35. Case control study
• Recruitment of:
▪ Case-patients affected with a disease
▪ Control-unaffected subjects
• Comparison of exposure status
• Observation of the past presence of one or
more potential risk factors
36. Cases Controls Total
Exposed a b -
Non exposed c d -
Total a+c b+d -
Odds Ratio= ad/bc
Presentation of the data of a case
control study in a 2 x 2 table
37. Strength of association between exposure & disease
Disease No disease
Exposure a b
Non- exposure c d
Odds (exposed) = Prob. of a case / prob. of non-case
= [a / (a+b)] / [b / (a+b)]
= a / b
Odds (not exposed) = c / d
Odds Ratio (OR) = (a / b) / (c / d)
= ad / bc
Odds of developing disease in the exposed group
(a/b) to the odds of developing disease in the
unexpected group (c/d).
38. Examples of case control studies
• Case control study to investigate an
outbreak
• Case control study to investigate the risk
factors for a rare disease
39. TOBACCO SMOKING AND PULMONARY TUBERCULOSIS
OBJECTIVE:
To study the association between tobacco smoking and pulmonary
tuberculosis using case-control methodology
40. METHODOLOGY
Definition of case: A bacillary TB case detected from the survey
belonging to the male sex and age group 20-50 years.
Definition of control: An individual examined and declared as
a non-case belonging to the male sex and age group 20-50 years.
Case: Control = 1:5
Measurement of exposure:
Using a questionnaire, following information was collected from
The study population:
1. Smoking status
2. Duration of smoking
3. Number smoked per day
41. Results:
Census status Cases Controls Total
Present* 85 462 547
Absent 3 29 32
Left 10 54 64
Dead 13 10 23
Fate unknown 1 1 2
Total 112 556 668
* This row only is considered for further analysis
43. ASSOCIATION BETWEEN TOBACCO SMOKING AND
PULMONARY TUBERCULOSIS
Case
category
Smoker Non-smoker
Case 64 21
Control 255 207
64 x 207
ODDS RATIO(OR) = ------------ = 2.47 ( 95% C.I: 1.42 to 4.34)
21 x 255 (P<0.001)
44. DOSE - RESPONSE RELATIONSHIP
Case
category
Smokers
Mild Moderate Heavy
(1-10/day) (11-20/day) (>20/day)
Non
smokers
Cases 25 21 18 21
Controls 140 66 49 207
Odds Ratio 1.76 3.14 3.62 -
Chi Square for linear trend = 17.946 (P<0.0001)
45. CUMULATIVE EFFECT OF SMOKING
Case
category
Smoking duration
<10 years 11-20 years >20 years
Non
smokers
Cases 14 22 28 21
Controls 81 89 85 207
Odds Ratio 1.70 2.44 3.25 -
Chi-Square for linear trend = 15.867 (P<0.0001)
46. CONCLUSION
1. There is an association between tobacco smoking and
pulmonary tuberculosis.
2. The association shows significant linear trend for dose
response relationship.
3. The association shows significant linear trend for cumulative
effect of smoking.
49. Take home messages
• Case control studies refine or test
hypotheses
• Case control studies come from cohorts
• Case and control definition are the keystone
of the design
• Exposure is collected retrospectively
50. CROSS SECTIONAL SURVEYS
TYPES OF QUESTIONS ADDRESSED
• What is the prevalence of hypertension in Chennai?
• What is the prevalence and distribution of known risk
factors for CHD in rural Tamil Nadu?
• How satisfied are patients attending government
hospitals in Chennai?
• What is the prevalence of TB in rural and urban
51. Cross-sectional survey
• Cross-sectional study involves examination of
a cross-section of the population (snapshot)
at one point of time. Individuals are
observed only once. It is usually based on a
random sample from a defined population.
Cross- sectional surveys are ideally suited to
study prevalence (point/period).
52. Potential objectives of a
cross sectional study
• Descriptive
▪ Estimate prevalence
• Correlation
Determine the Interrelationship among variables
• Analytic
▪ Compare the prevalence of a disease in various subgroups,
exposed and unexposed
▪ Compare the prevalence of an exposure in various
subgroups, affected and unaffected
53. Design of a Cross Sectional Study
Defined population
Gather data on exposure and disease
Exposed;
Have
disease
Exposed;
Do not
Have
Disease
Not
Exposed;
Have
Disease
Not
Exposed;
Do not
have
disease
54. Causality inference in
cross sectional studies
• Exposure and outcome examined at the same
time
• Prevalent cases
• Causal inference difficult
55. Prevalence of TB estimated from the three disease surveys
in Thiruvallur Dt. South India
Rounds of survey
Prevalence per 100,000
Smear- positive
irrespective of culture
status
Culture-positive
irrespective of smear
status
I survey (2000) 326 609
II survey (2002) 257 451
III survey (2005) 159 311
Annual decline %
(95% C. I)
12.3
(8.6-15.8)
12.6
(11.2-14.0)
TRC ICMR
56. Int J Tuberc Lung Dis 2008;12:916-20
Prevalence rates shown for 100,000 population
Prevalence of TB in Thiruvallur dt. South India
Consecutive community-based prevalence surveys
57. CROSS SECTIONAL SURVEYS
USES
- Assess prevalence of disease in a defined population
- Determine prevalence of risk factors of a disease
- Examine trends in disease or risk factors that can change
over time
- Provides quantitative estimates of the magnitude of a
health problem
- Generate hypotheses
- Measure health status of individuals in a defined population
- Plan health services and set priorities for disease control
59. CROSS SECTIONAL SURVEYS
LIMITATIONS
- Data on exposure to risk factors and
presence/absence of disease are collected
simultaneously
- Difficult to determine temporal relationship of
a presumed cause and effect
- Not useful to study disease etiology
- Not suitable for the study of rare diseases or
diseases with short duration
- Cannot measure disease incidence.
60. Limitations of causal inference in
analytical cross sectional studies
• Prevalent cases
• Exposure and outcome examined at the same
time
61. Take home messages
• Cohort studies go from exposure to outcome
• Case control studies go from outcome to
exposure
• Cross sectional studies look at outcome and
exposure at the same time
62. Expected Results and Interpretation
Remember to discuss expected results
and interpretation of analysis
Collection
of
Interpretation of
results to
answer...
Research
question
Study
design
Analysis of
findings
Close
the
loop
63. Summary: Successful
Research Design and Methods
• Bright idea
• Well developed and clearly described
methods
• Adequate sample size
• Reliable data collection
• Appropriate data analysis and interpretation
• Only minor limitations
• Clear pathway to strong conclusions
• Preparation of report