1. RESEARCH - STUDY DESIGNS,
DESCRIPTIVE STATISTICS
• Dr BIPUL BORTHAKUR
PROF Orthopaedics, SILCHAR ,ASSAM,INDIA
2. FORMULATION OF RESEARCH QUESTION
• forms backbone of a good research,
• building up of an appropriate hypothesis
• support a focused arguable thesis and construction of a logical
argument
3. • A good RQ is an asset as it:
1. Details the problem statement
2. Further describes and refines the issue under study
3. Adds focus to the problem statement
4. Guides data collection and analysis
5. Sets context of research.
4. A GOOD RESEARCH IS REPRESENTED BY
ACRONYM FINERMAPS
• Feasible
• Interesting
• Novel
• Ethical
• Relevant
• Manageable
5. FEASIBLE
• within the ability of the investigator to carry out. It should be backed
by an appropriate number of subjects and methodology as well as
time and funds to reach the conclusions.
• completed within the limited time and resources available to the
investigator.
• scope and scale of the project.
• collection of data and the proceedings of project can be completed
within the limited time and resources available to the investigator
6. INTERESTING
• with academic and intellectual debate
Novel
• should not simply copy questions investigated by other workers
• aim at confirming or refuting the already established findings,
establish new facts, or find new aspects of the established facts.
7. ETHICAL
• mandatory to get clearance from appropriate authorities
• minimizes the risk of harm to the participants
Manageable
It has the similar essence as of feasibility but mainly means that the
following research can be managed by the researcher
8. RELEVANT
• academic and intellectual interest to people
• It should establish a clear purpose for the research in relation to the
chosen field. For example, filling a gap in knowledge, analyzing
academic assumptions or professional practice, monitoring a
development in practice, comparing different approaches, or testing
theories within a specific population
9. APPROPRIATE
• RQ should be appropriate logically and scientifically for the
community and institution.
Systematic
Research is structured with specified steps to be taken in a specified
sequence in accordance with the well-defined set of rules though it
does not rule out creative thinking.
10. POTENTIAL VALUE AND PUBLISHABILITY
• research should aim for significant economic impact to reduce
unnecessary or excessive costs. Furthermore, the proposed study
should exist within a clinical, consumer, or policy-making context that
is amenable to evidence-based change.
11. HOW TO DEVELOP A RESEARCH QUESTION
• Begin by identifying a broader subject of interest that lends itself to investigate,
for example, hormone levels among hypospadias
• •Do preliminary research on the general topic to find out what research has
already been done and what literature already exists.[7] Therefore, one should
begin with "information gaps" (What do you already know about the problem?
For example, studies with results on testosterone levels among hypospadias
• •What do you still need to know? (e.g., levels of other reproductive hormones
among hypospadias)
• •What are the implied questions: The need to know about a problem will lead
to few implied questions. Each general question should lead to more specific
questions (e.g., how hormone levels differ among isolated hypospadias with
respect to that in normal population)
• •Narrow the scope and focus of research (e.g., assessment of reproductive
hormone levels among isolated hypospadias and hypospadias those with
associated anomalies)
• •Once question has been framed, one should evaluate it. This is to realize if
these would be effective RQs or if they need more revising
12. • • Is RQ clear? With so much research available on any given topic, RQs
must be as clear as possible in order to be effective in helping the writer
direct his or her research
• • Is the RQ focused? RQs must be specific enough to be well covered in
the space available
• • Is the RQ complex? RQs should not be answerable with a simple “yes”
or “no” or by easily found facts. They should, instead, require both
research and analysis on the part of the writer
• • Is the RQ one that is of interest to the researcher and potentially useful
to others? Is it a new issue or problem that needs to be solved or is it
attempting to shed light on previously researched topic
• • Is the RQ researchable? Consider the available time frame and the
required resources. Is the methodology to conduct the research
feasible?
• • Is the RQ measurable and will the process produce data that can be
supported or contradicted?
• • Is the RQ too broad or too narrow?
•
13. • •Create Hs: After formulating RQ, think where research is likely to be
progressing? What kind of argument is likely to be made/supported?
What would it mean if the research disputed the planned argument?
At this step, one can well be on the way to have a focus for the
research and construction of a thesis. Hs consists of more specific
predictions about the nature and direction of the relationship
between two variables. It is a predictive statement of the outcome of
research, dictate the method, and design of the research
• •Understand implications of your research: This is important for
application: whether one achieves to fill gap in knowledge and how
the results of the research have practical implications, for example,
to develop health policies or improve educational policies.
14. BRAINSTORM/CONCEPT MAP FOR
FORMULATING RESEARCH QUESTION
• First, identify what types of studies have been done in the past?
•Is there a unique area that is yet to be investigated or is there a
particular question that may be worth replicating?
•Begin to narrow the topic by asking open-ended “how” and “why”
questions
•Evaluate the question
•Develop a Hypothesis (Hs)
•Write down the RQ.
15. WRITING DOWN THE RESEARCH QUESTION
• State the question in your own words
• Write down the RQ as completely as possible. For example,
Evaluation of reproductive hormonal profile in children presenting
with isolated hypospadias)
•Divide your question into concepts. Narrow to two or three concepts
(reproductive hormonal profile, isolated hypospadias, compare with
normal/not isolated hypospadias–implied)
•Specify the population to be studied (children with isolated
hypospadias)
•Refer to the exposure or intervention to be investigated, if any
•Reflect the outcome of interest (hormonal profile).
16. HYPOTHESIS
• Once RQ is formulated, a Hs can be developed. Hs means
transformation of a RQ into an operational analog. It means a
statement as to what prediction one makes about the phenomenon
to be examined. More often, for case–control trial, null Hs is
generated which is later accepted or refuted.
• A strong Hs should have following characteristics:
• • Give insight into a RQ
• • Are testable and measurable by the proposed experiments
• Have logical basis
• Follows the most likely outcome, not the exceptional outcome.
17. EXAMPLES OF RESEARCH QUESTION AND
HYPOTHESIS
• Research question- 1 • Does reduced gap between the two segments
of the esophagus in patients of esophageal atresia reduces the
mortality and morbidity of such patients?
• Hypothesis-1 • Reduce d gap between the two segments of the
esophagus in patients of esophageal atresia reduces the mortality
and morbidity of such patients • In pediatric patients with
esophageal atresia, gap of <2 cm between two segments of the
esophagus and proper mobilization of proximal pouch reduces the
morbidity and mortality among such patients.
• Research question-2 • Does application of mitomycin C improves the
outcome in patient of corrosive esophageal strictures?
• Hypothesis-2 In patients aged 2–9 years with corrosive esophageal
strictures, 34 applications of mitomycin C in dosage of 0.4 mg/ml for
5 min over a period of 6 months improve the outcome in terms of
symptomatic and radiological relief.
18. STUDY DESIGN
• Design means whether study is cross-sectional or longitudinal.
• Cross-sectional – Study populations are examined once
• Longitudinal- Study populations are followed up at regular intervals
19. COMMON STUDY DESIGNS
1. CASE REPORT:
• <10 cases are presented
• Statistical tests cannot be employed
• Eg; Case report of one adolescent boy with repeated suicidal
attempts
20. 2.CASE SERIES
• >=10 cases with a particular disease are reported
• There is no control group
• Eg; case series of 32 silicosis patients.
• We can compare between two groups regarding duration of
symptoms, no. Of episodes etc by means of statistical tests
21. 3.ECOLOGICAL STUDY:
• Units of study are groups of people rather than individuals
• It is done when we have data only for the groups not individuals
• Eg; higher incidence of breast cancer in countries having high fat
consumption
• Doesnot demonstrate causal association
22. 4.CROSS SECTIONAL STUDY
• Population is examined only once , at one point of time
• Since there is no follow up only prevalence can be determined
• Helpful in assessing healthcare needs and during investigation of an
epidemic
• Eg; possible relationship of increased serum cholesterol level
[exposure] with ecg evidence of CHD [outcome]
23. 5.CASE CONTROL STUDY
• Retrospective study where both exposure [risk factor] and outcome
[disease] before start of the study.
• Proceeds from effect to cause
• Eg; to find out relationship between smoking and lung cancer , some
lung cancer patients as well as some non lung cancer patients are
included and asked for their smoking history
• This establishes association
• Controls should be selected by matching
24. 6. COHORT STUDY
• Exposure [smoking] has started or yet to start but outcome [lung
cancer] has not yet occured
• The cohort should be free from the disease under study
• Proceeds from cause to effect
• Use a nonexposed [nonsmoker] group to support or refute an
inference
25. TYPES OF COHORT STUDY
1. PROSPECTIVE COHORT STUDY
smokers and nonsmokers are identified , otherwise healthy and will
follow them for next 20-25 years
2. RETROSPECTIVE COHORT STUDY
both exposure and outcome have occured but direction of enquiry is
forward ie, cause to effect
26. 3. COMBINED PROSPECTIVE AND RETROSPECTIVE COHORT STUDY
Here cohort is identified from past records and assessed just like
retrospective cohort study and now after reaching the present year
the study is followed up in future, like prospective cohort study
27. • NESTED CASE CONTROL STUDY
1. It is a case control study done in a cohort study design.
2. The cases and controls are selected from a cohort population that
has been followed for a period of time.
3. Decreased chance of selection bias and recall bias
28. 7. RANDOMISED CONTROL TRIAL
• It is an experimental study with some deliberate manipulation done
by the researcher.
• Participants enrolled in the study are randomly assigned to one of
the following groups- intervention group & control group and
followed forward in time to determine if intervention has an impact
on a specific end outcome
29. RANDOMISATION
• A statistical procedure [by using a table of random numbers or by
computer generated random numbers] by which the participants are
allocated into study and control group so that every individual gets
an equal chance of being allocated into either group.
• Types of randomisation- simple, block, stratified, unequal
30. • Types of study design
1. Concurrent parallel study design- one group is exposed to specific
treatment and other group is not exposed. Patients remain in the
allocated group in the entire study period
2. Crossover study design- patients are assigned to study & control
groups by randomisation. Both are observed over time after
which the therapy is stopped till carryover effects of intervention
is nullified. Then the therapy is given to control groups.
31. • BLINDING
1. SINGLE BLIND TRIAL –only participant is blind
2. DOUBLE BLIND TRIAL- participant & investigator
3. TRIPLE BLIND TRIAL- participant, investigator , analysts are blind
32. FACTORIAL TRIAL
• More than one intervention in a single experiment is evaluated while
1]testing the treatment effects independently or 2] when the
treatments are thought to be complementary and the specific aim is
to investigate the treatment interactions
33. SYSTEMATIC REVIEW
• Collects and critically analyses multiple research studies
• Combination of large no. of good quality studies
• Steps –
1. framing question
2. search for review questions
3. assessment of quality
4. summarization of the evidence
5. interpretation of the findings
34. META-ANALYSIS
• Combines the results of individual studies to arrive at one overall
measure of the effect of a treatment or intervention.
• Provides precise estimate of treatment effect
• >= 2RCTs are needed
• Steps:
1. Search for articles
2. Assessment of publication bias
3. Assessment of quality of individual studies
35. • Presentation of findings
1. Forest plot
2. Heterogeneity
• Extent to which selected studies are different from each other
• Tested by : Cochrane’s Q & I^2 statistic[preferred]
• If heterogeneity is absent then analysis is done by fixed effects
modelling
• If heterogeneity is large analysis is done by meta -regression
36. •यं हि न व्यथयन्त्येते पुरुषं पुरुषषषभ |
समदु:खसुखं धीरं सोऽमृत्वाय कल्पते ||
15||
•yaṁ hi na vyathayantyete puruṣhaṁ
puruṣharṣhabha
sama-duḥkha-sukhaṁ dhīraṁ so
’mṛitatvāya kalpate
Meaning-O Arjun, noblest amongst men,
that person who is not affected by
happiness and distress, and remains steady
in both, becomes eligible for liberation.