RESEARCH METHODOLOGY & BIOSTATISTICS

Dr. Abhishek Kumbar

MD (Hom)

Data: The information or attributes are expressed in terms of numerical values.

There are two sources of data:
1) Primary source of data: A investigator directly collects the statistical data (or information) from informants.
(This is called first hand data). Methods of collection of primary data:
a. Direct personnel investigation method.
b. Indirect oral investigation method.
c. Mailed questionnaire method.
d. Schedules (or proforma) are filled by the investigators.
e. Laboratory method (or practical observation method).
2) Secondary source of data: The investigator collects the information from some agencies or some organizations
(government or private) which have been already collected by the organizations. (This is called second hand
data). Sources of secondary data:
a. Books
b. Journals
c. Research papers
d. Dissertations
e. Reviews and monographs6. Computerized search.
f. Government gazettes.
g. Official reports.
h. NGOs.
i. Universities etc.
Qualitative data: Also called as discrete data. It Quantitative data: Characterised by numerical values. This type
cannot be measured in numbers, but in extent of data can be measured and quantified in terms of numbers
and intensity. There are two types. They are and volume. There are two types. They are
Nominal data Ordinal data Interval data Ratio data
Data recognizing the Data differentiates Ordinal scale data placed with Data can place interval
qualitative according to some some meaningful interval between scale with some meaning
categories or groups. meaningful order. each successive data entry. ratio between it.
E.g.: male /female, E.g.: Ranking of medical E.g.: measurement of temperature E.g.: measurement of
white / black. students according to on Celsius scale. In this case the temperature on Kelvin
their examination interval between 900and 800 is scale (Kelvin scale an
performance. same as 400 and 300. absolute zero which is
absent in Celsius scale)

Measurements: proceeding from the nominal scale (the least precise type of scale) to ratio scale (the most
precise), relevant information is obtained increasingly
Nominal scale Ordinal scale Interval scale Ratio scale
∙ Nominal scale is ∙ Ordinal scales only permit ∙ Interval scales can have an ∙ Ratio scales have an
the least powerful the ranking of items from arbitrary zero, but it is not absolute or true zero
level of highest to lowest. possible to determine for of measurement. The
measurement. ∙ Ordinal measures have no them what may be called an term ‘absolute zero’ is
∙ It indicates no absolute values, and the absolute zero or the unique not as precise as it
order or distance real differences between origin. was once believed to
relationship and adjacent ranks may not be ∙ The primary limitation of the be.
has no arithmetic equal. interval scale is the lack of a ∙ Ratio scale represents
origin. ∙ All that can be said is that true zero; it does not have the the actual amounts of
∙ A nominal scale one person is higher or capacity to measure the variables. Measures
simply describes lower on the scale than complete absence of a trait or of physical
differences another, but more precise characteristic. dimensions such as
between things by comparisons cannot be ∙ The Fahrenheit scale is an weight, height,
assigning them to made. example of an interval scale. distance, etc. are
categories. examples.
 Variable: A variable is a quantity that varies and can take any one of a specified set of values. For example,
when collecting information on patient demographics, variables of interest may include gender, race or age.
As described by the psychologist Stanley Stevens in 1946, research data usually falls into one of the following
four types of variables:
Nominal variable
∙ Variables assessed on a
nominal scale are called
categorical variables.
∙ The order of the categories is
meaningless.
∙ The categories are mutually
exclusive and simply have
names.
∙ A special type of nominal
variable is a dichotomous
variable, which can take only
one of two values, for
example gender (male or
female). The data collected
are therefore binomial.
∙ If there are three or more
categories for a variable, the
data collected are
multinomial. For example, for
marital status, the categories
may be single, married,
divorced or widowed.
∙ Data collected for nominal
variables are usually
presented in the form of
contingency tables.
Ordinal variable
∙ An ordinal variable is
another type of
categorical variable.
When a ‘rank-ordered’
logical relationship exists
among the categories,
the variable is only then
known as an ordinal
variable.
∙ The categories may be
ranked in order of
magnitude.
∙ For example, there may
be ranked categories for
disease staging (none,
mild, moderate, severe)
or for a rating scale for
pain, whereby response
categories are assigned
numbers in the following
manner:
1. (no pain)
2. (mild pain)
3. (moderate pain)
4. (severe pain)
5. (unbearable pain)
Interval variable Ratio variable
∙ In addition to having all ∙ In addition to
the characteristics of having all the
nominal and ordinal characteristics of
variables, an interval interval variables,
variable is one where a ratio variable
the distance (or also has a natural
interval) between any zero point.
two categories is the ∙ Examples of ratio
same and constant. variables include:
∙ Examples of interval  height
variables include:  weight
o temperature, i.e. the  incidence or
difference between prevalence of
80 and 70 0F is the disease.
same as the
difference between
70 and 60 0F.
∙ Interval variables do
not have a natural zero
point. For example, in
the temperature
variable, there is no
natural zero, so we
cannot say that 40F is
twice as warm as 20F.
∙ On some occasions,
zero points are chosen
arbitrarily
 Research design
“a blueprint for conducting a study with maximum control over factors that may interfere with the validity of
the findings”.
Need for Research Design
∙ To determine frequency and burden of disease.
∙ To identify the risk factor.
∙ To determine the risk factor.
∙ To determine efficacy or effectiveness of new treatment.
∙ To evaluate community programme.

Features of Good Research Design

∙ There should be internal and external validity
∙ It should be reliable.
∙ It should be cost effective.
∙ It should have minimum errors and bias.
∙ It should be based on truth and fact.
∙ It should be capable of giving maximum information and knowledge.
 Ecological study:
“An ecological study is an observational study, in which the units of observation and analysis are at a group level,
rather than at an individual level.” Using aggregate data, they examine the association between exposures and
outcomes.
Steps for ecological study:
1. Data collection
∙ The data collected on the exposure and outcome variables should be at the same level of aggregation (e.g.
time period, city, region, country, continent).
∙ The studies often use data previously collected for an alternative purpose:
∙ Primary care data ∙ Infectious disease notification data
∙ Secondary care (hospital) data ∙ National survey data, e.g. population
∙ Mortality and census data registries or the census

2. Choosing correct level of measurements

∙ Aggregate measures: These combine data from individuals, summarised regionally or nationally, e.g. the
percentage of smokers in a city.
∙ Environmental measures: These are physical characteristics of a place in which members of each group work
or live, e.g. air-pollution level, temperature or climate of a country. Environmental measures have an
analogue at the individual level; however, they are not easy to measure. In other words, there may be
heterogeneity of the exposure level within groups.
∙ Global measures: These are attributes of places or groups for which there is no obvious individual analogue.
Examples include contextual variables such as laws restricting smoking in public places, the population
density of a country, or policies to improve equity in access to health care.
3. Choosing correct ecological study:
∙ Time trend studies - grouping is by time.
∙ Geographical studies - grouping is by place.
∙ Mixed design - combining a time trend and geographical study design.
4. Making inferences:
∙ Biologic inferences about effects on individual risks.
∙ Ecologic inferences about effects on group rates.
5. Interpreting the results: choosing the correct method to analyse the results of the 3 types of ecological study
designs that exists. Ex. Geographical study is most commonly used ecological study.

Ecological fallacy
∙ A limitation commonly faced when using ecological studies to make causal inferences between an exposure
and outcome, is ecological fallacy, which is a type of bias.
∙ “Ecological fallacy is the failure of an expected ecologic effect estimate to reflect the biologic effect that
exists at the individual level. In other words, ecological fallacy occurs when correlations based on grouped
data are incorrectly assumed to hold at the individual level.”

Advantages
∙ Can investigate whether differences in exposure
between areas are bigger than at the individual level.
∙ Utilise routinely collected health data; therefore, may
be relatively cheap and quick to conduct.
∙ Generate hypotheses which can be investigated at
the individual level using studies higher up in the
hierarchy of evidence.
∙ Can investigate the effect of exposures that are
measured over groups or areas, e.g. diet, air pollution
and temperature.
∙ Can search for associations in large populations.
Disadvantages
∙ Associations cannot be confirmed at the individual
level.
∙ There is potential for ecological fallacy when
applying grouped results to the individual level.
∙ There is lack of available data on confounding
factors.
∙ There is potential for systematic differences
between areas in how disease frequency is
measured, e.g. differences in disease classification
and coding.
∙ There is potential for systematic differences
between areas in how exposures are measured.
 Cross-sectional study:
“A cross-sectional study is a form of observational study that involves collecting data from a target population at
a single point in time.” This methodology is particularly useful for assessing the true burden of a disease or the
health needs of a population.
A cross-sectional study design is best suited when:
∙ Carrying out surveys on the prevalence of a disease (or other health-related characteristics) in order to assess its
burden in a defined population (and for planning the allocation of health resources accordingly).
∙ Carrying out surveys of views or attitudes, such as studies on smoking behaviour, alcohol consumption or patient
satisfaction.
∙ Studying the association between an exposure and disease onset for a chronic disease (where there is limited
information available on the time of onset of the disease).

Study design of
Cross-sectional study ∙ Cross-sectional studies are therefore useful
for planning the provision and allocation of
health resources. Most government surveys
conducted by the National Centre for Health
Statistics are cross-sectional studies.
∙ Cross-sectional studies are good at
discovering people with a disease who have not
previously sought any medical advice. Cross-
sectional studies are able to uncover the
iceberg of disease.

A. Descriptive cross-sectional studies: These studies are used to measure the prevalence and distribution of
a disease in a defined population.
B. Analytical cross-sectional studies: These studies are used to investigate the interrelationship between any
variables of interest. For example, a target population could be sampled to determine the characteristics
(age, sex, ethnicity, etc.) of those people with ischaemic heart disease.

Advantages
∙ Useful for measuring the prevalence of a disease in a
defined population and for planning the allocation of
health resources accordingly.
∙ Able to measure the prevalence of a disease for all
the exposure factors under investigation.
∙ Multiple exposures and outcomes can be measured
at the same time in the same cross-sectional study.
∙ Useful for generating hypotheses by studying the
association between exposure and disease onset for
chronic diseases.
∙ Generally quick (no long periods of follow-up) and
cheap to conduct.
∙ As routinely collected, readily available data are
commonly used, fewer resources are required to run
the study.
Disadvantages
∙ Usually cannot test epidemiologic
hypotheses as cross-sectional studies are
low on the hierarchy of evidence.
∙ Not useful for studying rare diseases or
diseases with a short duration.
∙ Cannot usually discern a temporal
relationship between an exposure and
disease outcome.
∙ The incidence rate of a disease outcome
cannot be estimated.
∙ Can be prone to selection bias due to poor
study participation response rates.
∙ Susceptible to measurement bias,
including survival and migration bias.
 Case-control study:
“A case–control study is a form of observational study that aims to identify risk factors for developing the
outcome of interest.”
Subjects with the outcome (cases) and without the outcome (controls) are selected and risk factor exposure
measurements are collected retrospectively in both groups either from the subject or from any available records.
Like cohort studies, case–control studies are useful when investigating causes of disease, or factors associated
with a particular condition. However, specifically, the case–control study design is best suited to answering
research questions investigating:
∙ The causes for an acute outbreak (e.g. Of legionnaire’s disease or gastroenteritis).
∙ The risk factors for a rare disease (e.g. Certain types of leukaemia). A prospective study, i.e. Cohort studies, would
take too long to identify a sufficient number of cases.
∙ The aetiology of a disease where there is a long time lag between an exposure and disease outcome.

Study design

Exposure odds ratio = Odds of exposure in cases

Odds of exposure in controls
= d1/d0 = d1 x h0
h1/h0 d0 x h1

Advantages
Efficient for studying the effect of exposures on rare
diseases.
Efficient for studying the effect of exposures on
disease outcomes with long latency periods.
Useful for studying the effect of multiple exposures
on disease outcome.
A retrospective study; therefore there are no long
periods of follow-up as the investigator does not
need to wait for incident cases.
As cases are identified at the beginning of the study,
there is no loss to follow-up.
Relatively quick, cheap and easy to perform.
Disadvantages
Reverse causality may be an issue, as the temporal
sequence of exposure and outcome was not observed.
A retrospective study, therefore particularly prone to recall
bias and interviewer bias.
Can be prone to selection bias when choosing cases or
controls.
Limited to investigating the effect of exposures on only
one outcome.
Not suitable when investigating the effect of rare
exposures on disease outcome.
The incidence rate of disease outcome cannot be
estimated (unless the case–control study is population
based).
 Cohort study:

“A cohort study is a form of observational study that aims to investigate whether exposure of subjects to a certain
aetiological factor will affect the incidence of a disease in the future.”
Most cohort studies assess the harm or benefit of a risk factor, especially if little is known about the effect of this
exposure. Specifically, the cohort study design is best suited to answering research questions:
∙ where the exposure is rare.
∙ where it would be impossible, or unethical, to expose subjects deliberately to the risk factor (e.g. if
investigating the effects of smoking or alcohol).
∙ where the number of subjects needed to detect an effect are too great for a randomised controlled trial to be
feasible. However, note that when a disease is very rare, the numbers needed may be too large for a
prospective cohort study, and a retrospective cohort study design or a case–control study is preferred.

Study design Prospective and retrospective cohort study

In a retrospective study design:

∙ If the exposure to risk has already occurred
∙ If there is a long time lag between exposure and
disease outcome.
∙ If the disease outcome is rare.

In a prospective study design:

∙ If there is a short time lag between exposure and
disease outcome.

1. Risk of disease = Number of new cases of disease

over study period
Total number of subjects initially
disease-free
Risk of disease = (d1 + d0)
n
2. Risk ratio = risk in exposed group
d = disease h = healthy risk in unexposed group
Risk ratio = d1 / (d1 + h1)
1 = exposed 0 = unexposed d1 / (d1 + h1)
3. Risk difference
= risk in exposed group - risk in unexposed group
= [d1 / (d1 + h1)] – [d1 / (d1 + h1)]
 Advantages Disadvantages
Prospective Prospective
∙ The time sequence between exposure and disease is ∙ As the length of follow-up increases, the study is
clear as exposure status is measured before disease more prone to selection bias as subjects migrate or
onset (preventing reverse causality). leave the study.
∙ The exposure can be measured at various time points, ∙ Following up subjects for disease occurrence is time
thus establishing any changes in exposure status during consuming, therefore costly.
follow-up. ∙ Not suitable for rare diseases as the length of follow-
∙ Exposure is measured before disease onset, thus up may be considerable to get enough incident cases
minimising bias in exposure measurement. of the disease outcome.
∙ Not suitable for rare diseases as very large sample
sizes are required.
∙ Prone to surveillance bias.
∙ Unsuitable if there is a long time lag between
exposure and outcome.
Retrospective Retrospective
∙ It is possible to measure long-term effects of exposure ∙ Prone to recall bias (unlike prospective cohort
as the follow-up period has already occurred. studies).
∙ Useful for rare outcomes (diseases with low incidence), ∙ Reverse causality may be an issue, as the temporal
as the outcome has already happened. sequence of exposure and outcome was not
∙ Less expensive than prospective cohort studies because observed.
outcome and exposure have already occurred.
∙ Useful if there is a long time lag between exposure and
outcome.
Prospective and Retrospective Prospective and Retrospective
∙ Information on exposure to a wide range of factors can ∙ Disease outcome or the aetiology of disease may
be assessed (cohort studies are useful if little is known change over time.
about the effect of an exposure). ∙ Prone to interviewer bias
∙ They can provide information on associations between
an exposure and multiple outcomes (retrospective
studies are better for this).
∙ It is possible to directly measure the incidence/risk of
disease in the exposed and unexposed groups.
∙ It is possible to investigate the effect of rare exposures
(frequency less than 20%) on disease outcome
 Randomised controlled trial or Randomised clinical trial

“A randomised control trail is an interventional study during which study participants are randomised to
different treatment options.”
Study design

“Randomised clinical trial aim to make a fair comparison between a new treatment and the existing
treatment, or between two (or more) existing treatments, to see which one works best”

The essential steps involved in a RCT are:

1. Formulate the hypothesis
2. Define the methods of recruitment, including the inclusion and exclusion criteria.
3. Define the intervention.
4. Define the comparison group. (new treatment and placebo treatment)
5. Determine the sample size.
6. Specify the outcome measures that will be used to assess the effectiveness of the intervention.
7. Obtain ethical approval.
8. Obtain informed consent before the study participants are randomised to either the intervention or control.
9. Generate and conceal an allocation sequence to ensure randomisation.
10. Indicate whether the assessors and/or study participants have any knowledge of the treatment allocation
(blinding).
11. Perform an intention to treat analysis.

Blinding
∙ Blinding refers to patients and investigators (including those involved in recruitment and assessing the
outcome) having no knowledge of treatment allocation.
∙ Traditionally, blinded RCTs have been classified as ‘single-blind’, ‘double-blind’ or ‘triple-blind’. However,
due to inconsistency in definitions of these terms and lack of clarity in journals, it is better to specify who
exactly was blinded and how.
∙ If the intervention is an active drug, it is possible for both the subject and investigator to be blind to
treatment allocation if the comparison group takes an inactive placebo, which looks, tastes and feels exactly
like the active drug.
∙ RCTs may also use an active placebo that mimics the common side effects of the drug under study.
∙ It is important to note that blinding may not be possible if the RCT involves:
∙ a technology, e.g. surgery versus chemotherapy.
∙ a programme of care, e.g. exercise therapy versus medication.
∙ In these studies, known as open-label trials, randomisation should still be used and the outcome assessor
should still be blind (if possible) to which treatment the participant received.
 Cluster / Community randomised trials (CRTs) involve randomisation of groups (clusters) of individuals to
control or intervention conditions. The CRT design is commonly used to evaluate non-drug interventions, such as
policy and service delivery interventions. Its use is likely to grow as we move towards the learning healthcare
system and large simple trials.

Randomised Control Trails

Advantages
Provides the strongest evidence of any study
design to determine the effectiveness or safety
of a given intervention.
Has the most rigorous study design to
determine whether a cause–effect relationship
exists between an intervention and outcome
(there is a clear temporal sequence – exposure
precedes outcome).
Prospective study design, therefore able to
measure disease risk.
Enables blinding; therefore, bias is minimised.
Randomisation can control for known or
unknown confounders.
Disadvantages
Limitations to external validity due to:
∙ Strict eligibility criteria regarding patient characteristics
∙ Recording unrepresentative outcome measures
∙ Using difficult study procedures.
Difficult to detect small effects as this would require a very
large sample size. This may be an issue when investigating:
∙ Rare outcomes (e.g. sudden infant death syndrome)
∙ Uncommon adverse outcomes (e.g. a rare side effect of a
drug)
Costly to study late outcomes which require long
periods of follow-up
Relatively expensive
The efficacy of the intervention may be different under trial
conditions, where protocols are strictly followed, compared
to normal practice
 Hybrid / Mixed research method
“Mixed methods research is the type of research in which a researcher or team of researchers combines
elements of qualitative and quantitative research approaches (e. g., use of qualitative and quantitative viewpoints,
data collection, analysis, inference techniques) for the broad purposes of breadth and depth of understanding and
corroboration.”
A study design that takes a dual focus in assessing clinical effectiveness and implementation. Hybrid designs can
typically take 1 of 3 approaches:
a. testing effects of a clinical intervention on relevant outcomes while observing and gathering information on
implementation;
b. dual testing of clinical and implementation interventions/strategies;
c. testing of an implementation strategy while observing and gathering information on the clinical
intervention’s impact on relevant outcomes. Such dual foci are always stated a priori.

Hybrid Design Characteristics and Key Challenges

Study
Hybrid Trial Type 1
Characteristic
Research aims Primary aim: determine
effectiveness of a clinical
intervention
Secondary aim: better
understand context for
implementation
Research Primary question: will a
questions clinical treatment work in
(examples) this setting/these patients?
Secondary question: what
are potential
barriers/facilitators to a
treatment’s widespread
implementation?
Units of Patient, clinical unit
randomization
Comparison Placebo, treatment as usual,
conditions competing treatment
Sampling Patient: limited restrictions,
frames but some
inclusion/exclusion criteria
Provider, clinical unit,
facility, system: choose
subsample from relevant
participants.
Evaluation Primary aim: quantitative,
methods summative
Secondary aim: mixed
methods, qualitative,
process-oriented, could also
inform interpretation of
primary aim findings
Measures Primary aim: patient
symptoms and functioning,
possibly cost
Secondary aim: feasibility
and acceptability of
implementing clinical
treatment, sustainability
potential, barriers and
facilitators to
implementation
Hybrid Trial Type 2
Co-primary aim: determine effectiveness of
a clinical intervention
Co-primary aim: determine feasibility and
potential utility of an implementation
intervention/strategy
Co-primary question: will a clinical
treatment work in this setting/these
patients?
Co-primary question: does the
implementation method show promise
(either alone or in comparison with another
method) in facilitating implementation of a
clinical treatment?
Clinical effectiveness: see type I
Implementation: see type III, although may
be nonrandomized, for example, case study
Clinical effectiveness: see type I
Implementation: see type III, although may
be nonrandomized, for example, case study
Patient: limited restrictions, but some
inclusion/exclusion criteria
Providers/clinics/facility/systems; consider
“optimal” cases
Clinical effectiveness aim: quantitative,
summative
Implementation aim: mixed method;
quantitative, qualitative; formative and
summative
Clinical effectiveness aim: patient
symptoms and functioning, possibly cost
effectiveness
Implementation aim: adoption of clinical
treatment and fidelity to it, as well as
related factors
Hybrid Trial Type 3
Primary aim: determine utility of an
implementation intervention/strategy
Secondary aim: assess clinical outcomes
associated with implementation trial
Primary question: which method works
better in facilitating implementation of a
clinical treatment?
Secondary question: are clinical outcomes
acceptable?
Provider, clinical unit, facility, system
Provider, clinical unit, facility, system:
implementation as usual, competing
implementation strategy
Provider/clinic/facility/system: either
“optimal” cases or a more heterogeneous
group
Secondary: all or selected patients included
in study locations
Primary aim: mixed-method, quantitative,
qualitative, formative, and summative
Secondary aim: quantitative, summative
Primary aim: adoption of clinical treatment
and fidelity to it, as well as related factors
Secondary aim: patient symptoms,
functioning, services use
 Study
Hybrid Trial Type 1 Hybrid Trial Type 2 Hybrid Trial Type 3
Characteristic
Potential Generating “buy in” among Generating “buy in” among implementation Primary data collection with patients in
design clinical researchers for researchers for clinical intervention aims large, multisite implementation trials can
challenges implementation aims These studies will require more research be unfeasible, and studies might need to
Insuring appropriate expertise and personnel, as well as larger rely on subsamples of patients, medical
expertise on study team to budgets, than non-hybrids record review, and/or administrative data.
conduct rigorous Secondary Insuring appropriate expertise on study Patient outcomes data will not be as
aim team to rigorously conduct both aims extensive as in traditional effectiveness
These studies will likely “Creep” of clinical treatment away from trials or even other Hybrid types, and might
require more research fidelity needed for optimal effectiveness be insufficient to answer some questions
expertise and personnel, IRB complexities with multiple types of “Creep” of clinical treatment away from
and larger budgets, than participants fidelity needed for optimal effectiveness
non-hybrids IRB complexities with multiple types of
participants

Compare and contrast qualitative and quantitative research

Quantitative Research Qualitative Research
A type of educational research in which the A type of educational research in which the researcher
researcher decides what to study. relies on the views of the participants.
Characteristics
∙ Ask specific narrow Qs. ∙ Ask broad, general Qs.
∙ Collects data from participants. ∙ Collecting data consisting largely of words (text) or
∙ Analyses numbers using statistics. image (picture).
∙ Conducts the inquiry in unbiased, objective ∙ Descriptions and analysis of words for themes.
manner. ∙ Conducts inquiry in subjective, biased manner
Differences
Description of trends or an explanation of ∙ An exploration in which little is known about the
variables’ relationships. problem.
∙ A detailed understanding of a central phenomenon.
Reviewing the literature
∙ Major role through suggesting the RQ to be asked. ∙ Minor role in suggesting SRQ to be asked.
∙ Justifying the R problem and the need for the ∙ Justify the importance of studying the research problem
direction of the study.
Purpose
∙ Be specific and narrow. ∙ Be general and broad.
∙ Seek measurable, observable data on variables ∙ Seek to understand the participants’ experiences.
Collecting data
∙ Collecting data using instruments with preset Qs ∙ Collecting data using forms with general, emerging Qs to
and Res. permit the participant to generate responses.
∙ Collecting info from a large number of individuals. ∙ Gathering word(text) or image(picture) data.
∙ Collecting info from a small number of individuals or
sites.
Analysing and Interpreting data
∙ Data analysis tends to consist of statistical analysis. ∙ Text analysis.
∙ Describing trends, comparing group differences, ∙ A description of themes.
relating variables. ∙ Stating the larger meaning of findings.
∙ Interpretation tends to consist of comparing results
with prior predictions and past research.
Reporting and Evaluating research
∙ Tend to use standard fixed structure and ∙ A flexible, emerging structure and evaluative criteria.
evaluation criteria. ∙ Take a subjective and biased approach.
∙ Take an objective and unbiased approach.
 Conceptual research
∙ Conceptual research is that related to some
abstract idea(s) or theory.
∙ It is generally used by philosophers and
thinkers to develop new concepts or to
reinterpret existing ones.
Empirical research
∙ Empirical research relies on experience or observation alone,
often without due regard for system and theory.
∙ It is data-based research, coming up with conclusions which
are capable of being verified by observation or experiment.
∙ We can also call it as experimental type of research.
∙ Empirical research is appropriate when proof is sought that
certain variables affect other variables in some way.
∙ Evidence gathered through experiments or empirical studies
is today considered to be the most powerful support possible
for a given hypothesis.

Action research
Practical action research
“Action research is focused on the immediate application and not on the
development of theory.”

∙ It has placed its emphasis on a real problem in a local setting.

∙ Its findings are to be evaluated in terms of local applicability, not in terms of
universal validity.
∙ Provide means to improve the practices.
∙ Promote a process of testing new ideas.

Operational research
“Operations Research (OR) is the study of mathematical models for complex organizational systems.”

Steps used in operational research

∙ It makes sense to make the best use of available resources.

∙ Today's global markets and instant communications mean
that customers expect high-quality products and services
when they need them, where they need them.
∙ Organizations, whether public or private, need to provide
these products and services as effectively and efficiently as
possible.
∙ This requires careful planning and analysis the hallmarks of
good OR.
∙ This is usually based on process modelling, analysis of
options or business analytics.
 Systematic review
Systematic reviews are:
1. Required to establish the clinical benefit and cost-effectiveness of an intervention.
2. Used by the National Institute for Health and Clinical Excellence to appraise single or multiple interventions.
3. Crucial when there is an important, potentially lifesaving, question which has been addressed by a number of
primary randomised controlled trials but there is still uncertainty about the answer.

Advantages
∙ Appear at the top of the ‘hierarchy of evidence’ that informs
evidence-based practice, thus giving us the best possible
estimate of any true effect.
∙ Can shorten the time lag between research practice and the
implementation of new findings into clinical practice.
∙ Relatively quicker and less costly to perform than a new study.
∙ Large amounts of information are critically appraised and
synthesised in order to reduce errors (including bias) and
improve the accuracy and reliability of the findings.
∙ Compared to a single study, the results can often be generalised
to a broader population across a wide range of settings.
∙ If the studies included in the review give consistent results, it
provides evidence that the phenomenon is robust and
transferrable.
∙ If the studies included in the review give inconsistent results, any
sources of variation can be studied.
∙ A meta-analysis has high power to detect exposure effects and
estimate these effects with greater precision than single studies.
Disadvantages
∙ Require considerably more effort than
traditional reviews.
∙ The clinical questions posed are often too
narrow, thus reducing the applicability of
the findings to your patient.
∙ Sometimes the interventions reviewed do
not reflect current practice.
∙ There may be an insufficient number of
high-quality studies available for review.
∙ The underlying physiological effects of an
intervention are not considered.
∙ Systematic reviews rarely consider the fact
that some interventions are delivered as
part of a larger package of care.

Hypothesis
“Hypothesis is a prediction or explanation of the relationship between two or more variables.”
It should be based on logic and science.

Steps involved in testing hypothesis

Steps involved in hypothesis testing are,
1. Specify the null hypothesis and the alternative hypothesis.
2. Collect the data and determine, what statistical test is appropriate for data analysis.
3. Perform the statistical test, to compute the P-value.
4. Use the P-value, to make a decision in favour of the null or alternative hypothesis.

Type I and II Errors

Type I Error: Incorrectly reject the null hypothesis when it is true.
∙ When random sampling causes your data to show a statistically significant association/difference, but there
really is no effect, a type I error has been made.
∙ Your conclusion that the sample means of the two groups are really associated/different is incorrect.
∙ Type I error is also our alpha (a) level in our hypothesis test, which represents the level of error we are willing to
accept in our study.

Type II Error: Incorrectly fail to reject the null hypothesis when it is false.
∙ When random sampling causes your data to not show a statistically significant association/difference, but there
really is an effect, a type II error has been made.
∙ Your conclusion that the sample means of the two groups are not really associated/different is incorrect.
∙ We use the letter beta (b) to represent type II error.
 Ethics in Clinical Research
Some of the influential codes of ethics and regulations that guide ethical clinical research include:
∙ Nuremberg Code (1947)
∙ Declaration of Helsinki (2000)
∙ Belmont Report (1979)
∙ CIOMS (2002)
∙ U.S. Common Rule (1991)
Using these sources of guidance and others, seven main principles have been described as guiding the conduct
of ethical research:
1. Social and clinical value: In other words, answers to the research question should contribute to scientific
understanding of health or improve our ways of preventing, treating, or caring for people with a given disease.
Only if society will gain useful knowledge — which requires sharing results, both negative and positive — can
exposing human subjects to the risk and burden of research be justified.
2. Scientific validity: A study should be designed in a way that will get an understandable answer to the valuable
research question.
3. Fair subject selection: The primary basis for recruiting and enrolling groups and individuals should be the
scientific goals of the study — not vulnerability, privilege, or other factors unrelated to the purposes of the
study.
4. Favourable risk-benefit ratio: Uncertainty about the degree of risks and benefits associated with a drug, device,
or procedure being tested is inherent in clinical research — otherwise there would be little point to doing the
research.
5. Independent review: To minimize potential conflicts of interest and make sure a study is ethically acceptable
before it even starts, an independent review panel with no vested interest in the particular study should review
the proposal and ask important questions
6. Informed consent: For research to be ethical, most agree that individuals should make their own decision about
whether they want to participate or continue participating in research.
7. Respect for potential and enrolled subjects: Individuals should be treated with respect from the time they are
approached for possible participation—even if they refuse enrolment in a study—throughout their participation
and after their participation ends. This includes,
a. Respecting their privacy and keeping their private information confidential.
b. Respecting their right to change their mind, to decide that the research does not match their interests, and
to withdraw without penalty.
c. Informing them of new information that might emerge in the course of research, which might change their
assessment of the risks and benefits of participating.
d. Monitoring their welfare and, if they experience adverse reactions, untoward events, or changes in clinical
status, ensuring appropriate treatment and, when necessary, removal from the study.
e. Informing them about what was learned from the research. Most researchers do a good job of monitoring
the volunteers’ welfare and making sure they are okay. They are not always so good about distributing the
study results. If they don’t tell you, ask.
 Measures of dispersion / variation

1. Range: Simplest measure of variation.

Difference between the largest and the smallest values in a set of data:
Range = Xlargest – Xsmallest

Disadvantages of the Range:

∙ Ignores the way in which data are distributed.
∙ Sensitive to outliers.
2. Quartiles: Quartiles split the ranked data into 4 segments with an equal number of values per segment.
3. Coefficient of Range: Coefficient of Range = [L-S]/[L+S]
Where,
L = Largest value in the data set
S = Smallest value in the data set
4. Interquartile Range: Can eliminate some outlier problems by using the interquartile range.
Eliminate some high- and low-valued observations and calculate the range from the remaining values
Interquartile range = 3rd quartile – 1st quartile
= Q3 – Q1
5. Variance: Average (approximately) of squared deviations of values from the mean.
6. Standard Deviation:
Standard deviation is defined as the square-root of the
average of squares of deviations, when such deviations for the
values of individual items in a series are obtained from the
arithmetic average.

Example:
7. Mean deviation: is the average of difference of the values of items from some average of the series.

Such a difference is technically described as deviation.

In calculating mean deviation, we ignore the minus sign

of deviations while taking their total for obtaining the
mean deviation.