INTRODUCTION AND NEED FOR EPIDEMIOLOGICAL DATA IN CLINICAL

RESEARCH
DR. S.N.MANJUNATHA MD, DNB, DPHA, DHM, DMLE

The plural of anecdote is not data

Research is the aspiration of the intellect of the individual and the collective to gather
evidence to unravel not just the mysteries beyond but also ‘rediscover the truth and the
fact’ which may be obvious.
The reasons for undertaking research can be many: it can be for
a) Academic requirements – doing a thesis or dissertation, submitting an assignment,
doing a project work.
b) Participating in a scientific meet; National / Sub national conference.
c) Career prospects – doing a paper, impact of the published paper.
d) Peer pressure – to answer an oft question –are you not doing it? Jealousy!
e) Planning for health care of the community either by Government or Non-
government agencies – commissioned studies, evaluation studies.
f) Miscellaneous: Research for the sake of research.
The purposes of research itself can be very different. It could be for:
a) Documentation
b) Exploratory
c) Evaluation
d) Estimation
e) Repeat the research in ‘our context’

In recent years, the quantitative approach in clinical research and clinical practice has
become increasingly important. Epidemiology, clinical decision analysis and biostatistics
are the disciplines traditionally involved in the quantitative aspects of medical problems.
In daily practice modern clinicians are increasingly expected to combine patient care with
clinical research. There is a growing need for physicians who can both initiate research
and participate in it.
Clinical Epidemiology – which is empirical and patient oriented – is most suited for such
a combination. It involves diagnostic, prognostic, preventive and therapeutic studies that
follow directly from quantitative study designs.

Clinical research directs two basic questions: How can we provide the most accurate
diagnosis? And how can we give the most effective treatment to our patients?
Patients are individuals who differ from each other in many aspects; this may influence
the effects of therapy. At the same time, one diagnosis often covers many diseases, which
also may differ in degree of severity, response to therapy, etc. Clinicians therefore have
to perform clinical trials and test methods to distinguish subgroups of patients and
diseases. The ultimate goal of clinical research is a tailor-made diagnosis and therapy for
the individual patient.
EPIDEMIOLOGICAL RESEARCH
Epidemiology is the study of the distribution and determinants of health-related states or
events in specified populations, and the application of the study to control health
problems. “Study” includes surveillance, observation, hypothesis testing, analytic
research and experiments. “Distribution” refers to analysis by time, place and class of
persons affected. “Determinants” are all the physical, biological, social, cross-cultural
and behavioural factors that influence health. “Health related states and events” include
diseases, causes of death, behaviours such as use of tobacco, reactions to preventative
regimens and provision and use of health services. “Specified populations” are those with
identifiable characteristics such as precisely defined numbers. “Application of study to
control health problems” makes explicit the aim of epidemiology - to promote, protect
and restore health.” Epidemiological research is thus concerned, through the collection of
data related to health, with the description of health and welfare in populations, and with
the goal of improving health. Some epidemiological research may require a study of
entire populations and go beyond an individual institution or organisation.
Epidemiological research is part of wider public health and health services’ research. It is
concerned with improvements of health and welfare in human populations and with
improving the efficiency and performance of human health services. Public health and
health services’ research are often carried out with human participants, or data or
biological samples from them, and such research provides important new knowledge that
is not readily obtainable in other ways.
In epidemiological research, medically relevant information about individuals and groups
is accumulated. Features of groups of persons may be investigated whether or not the
information was originally obtained for research purposes.
Different methods are used in carrying out an epidemiological investigation:
Surveillance and descriptive studies are used to study distribution; analytic studies are
used to study determinants (causes, risk factors).
Surveillance and Surveys
Surveillance may be defined as the process of systematic collection, orderly
consolidation, and analysis of data, with prompt dissemination and feedback of the results
to those who need to know, particularly those who are in a position to take action.
Surveillance generally uses methods distinguished by their practicality, uniformity and
rapidity rather than by accuracy or completeness. It is usually based on information
collected as part of routine health care, although it may sometimes be based on repeated
purposive surveys.
A survey may be defined as an investigation in which information is systematically
collected. The term survey is sometimes used in a narrow sense to refer specifically to a
“field survey”.
Uses of surveillance
_ Determine trends over time.
_ Gather simple information on determinants of risk or other geographical subdivisions,
disease, geographical subunits or categories of individuals at risk (districts or age groups
with higher rates).
_ Detect the occurrence of disease (sporadic, endemic, epidemic).
_ Set goals and targets based on information regarding prevalence and trends in order to
design health interventions.
_ Assess whether health goals and targets are being reached.
Descriptive studies
Descriptive studies may be defined as studies that describe the patterns of disease
occurrence by time, place, and person.
Uses of descriptive studies
_ In health planning and administration; descriptive studies and the analysis of their
results allow planners and administrators to allocate resources efficiently.
_ They are also used for hypothesis generation; often providing first important clues
about etiology.
Types of descriptive studies
Case reports or case series
_ These describe socio-demographic, behavioural and/or medical characteristics for one
or more persons with a similar diagnosis (example: characteristics of children admitted to
a hospital with cerebral malaria during a one-year period).
_ They provide an important link between clinical medicine and epidemiology.
_ They are often useful for generating hypotheses and examining new diseases.
However, conclusions about aetiology or risk factors cannot be made without having
undertaken analytic studies to examine the expected frequency of exposure to the
aetiological or risk factor in a group that does not have the illness under investigation.
Ecological Studies
_ These may compare disease frequencies among different groups during the same
period, or compare disease frequencies in the same population at different points in time
as a function of some exposure.
_ Ecological studies usually are quick and easy to perform, and can be undertaken with
already available information, but great care is needed to avoid reaching conclusions
based on spurious associations.
_ Ecological studies cannot link exposure to outcome in a given individual.
_ Descriptive studies constitute one of the first steps in outbreak investigation; and should
always be undertaken before initiating further analytic studies
Analytic studies
Analytic studies may be defined as studies used to test hypotheses concerning the
relationship between a suspected risk factor and an outcome, and to measure the
magnitude of the association effect, and its statistical significance. An analytic study
always implies a comparison among two or more groups.
There are two main types of analytic studies: observational and experimental
Observational studies
_ Most analytic studies fall in this category.
_There is no human intervention involved in assigning study groups; one simply observes
the relationship between exposure and disease.
_ Observational studies are subject to many potential biases. Careful design and analysis
may help avoid many of these biases.
_ There are three basic categories of observational studies: cross-sectional, cohort studies
and case-control studies
Cross-sectional Studies (Surveys)
_ These examine the relationship between a disease or other health-related characteristic
and other variables of interest they exist in a population at a given time. The presence or
absence (or the level) of a characteristic is examined in each member of the study
population or in a representative sample. These studies are used to obtain information not
routinely available from surveillance or case series.
_ Cross-sectional studies provide no information on the temporal sequence of cause and
effect. In surveys examining the association between an exposure and an outcome, both
are measured simultaneously and it is often hard to determine whether the exposure
preceded the outcome or vice versa.
_ Surveys may simply describe characteristics or behaviours within a study population
(autism prevalence, vaccine coverage); or may be used to examine potential risk factors
(e.g., how those who receive vaccination differ from those who do not).
_ In general, surveys measure the situation at a given moment – prevalence – rather
than the occurrence of new events – incidence.
Surveys are not recommended for the study of rare diseases or of diseases with a short
duration, nor are they suited to the study of rare exposures.
Case Control Studies
Case-Control Studies proceed conceptually from outcome to exposure; start with groups
affected with the outcome (in the case of a disease, the “ill” group) and groups not
affected (“well”) and retrospectively determine the rates of exposure to a risk factor for
each group to see if these rates differed.
_ Group of subjects with the disease or other outcome variable (cases), and group of
subjects without the disease or other outcome variable (controls) are identified.
Information about previous exposures is obtained for cases and controls, and the
frequency of exposure is compared for the two groups.
_ In case-control studies, both exposure and disease are normally considered to have
occurred prior to enrolment in the study.
Cohort Studies
Cohort studies proceed conceptually from exposure to outcome; starting with exposed
and unexposed groups and following them to see if the rates of occurrence of the
outcome in the two groups differ.
_ Study groups are identified by exposure status prior to ascertainment of their disease
status; both exposed and unexposed groups are then followed prospectively in an
identical manner until they develop the disease under study, until the study ends, or the
subjects die or are lost to follow-up. Both cohorts should have similar characteristics
except for the exposure under investigation.
_ Cohort studies differ from experimental studies in that the investigator does not
determine exposure status. This is determined by genetics or biology (sex, presence or
absence of genetic disease, etc.), subject's choice (smoking behaviour, use of
contraceptives, sexual behaviours, food consumption, etc.) or other circumstances (rural
versus urban, socio-economic status etc.).
_ In some studies, called retrospective cohort studies, exposure and outcome both lie in
the past (before enrolment). The main conceptual element to remember is that the
retrospective cohort proceeds from exposure to disease.
Experimental studies
_ The person conducting the study randomizes the subjects into exposed and unexposed
groups and follows them over time to compare their rates of disease development.
Examples may include trials of the efficacy of a new drug compared with the efficacy of
the drug currently in use.
_ Randomization helps ensure comparability of the groups and avoids many of the biases
inherent in non-experimental studies; for this reason experimental study have been
considered as a widely accepted "gold-standard".
_ Experimental studies are nevertheless expensive; they are not suitable for the study of
rare disease outcomes, may take a long time to perform, often present complex problems
of ethics, or may simply not be feasible (e.g. randomized trials of the health benefits of
breastfeeding). They may also provide results different from those observed under field
conditions.
The logical sequence of epidemiological studies
In epidemiological research, the current state of knowledge often determines the most
logical study design. One usually sees a progression from hypothesis-generating to
hypothesis-testing studies. For example, hypotheses are often generated by methods such
as surveillance, case reports, case series, or ecological studies. These hypotheses are then
tested using data from experience, from previous cross-sectional studies, from case-
control studies, or from retrospective cohort studies, which can be done relatively quickly
and cheaply. If these studies lend support to the hypothesis, a prospective cohort study
may be undertaken. Finally, in some situations, a randomized clinical trial may be
undertaken.

DATA MANAGEMENT IN EPIDEMIOLOGICAL RESEARCH

Data management involves data collection, entry, verification and data retrieval when
required. The aim of data management is to turn information from the subject into a
report, efficiently and without errors. Data management is the most important aspect of
any epidemiological research. The investigator is responsible for the collection, quality,
recording, maintenance and retrieval of the data arising from the epidemiological study.
All steps involved in data management should be documented properly. In any later
auditing of the study by an independent authority, the documentation will allow them to
perform a step-by-step retrospective assessment of data quality and study performance.
Before starting any epidemiological projects a " standard operating (SOP)" Document
should be prepared. The SOP provides detailed written instructions for procedures and a
general framework for the implementation of all the function and activities of each
procedure.
Once the protocol is finalized, a sample of subjects should be selected and a trial run
should be carried out in the field. This trial run will help the investigator and personnel
involved in the epidemiological research to get a feel of the study. The data for the trial
cases should be entered in the prototype of the case record forms. The person entering the
data must be trained to enter the data properly and understand the significance of all the
information collected. If there are any clarifications by the staff in the data entry, it
should be done at the beginning of the study to avoid future confusion. Once the trial fun
is over, any problems with data collection should be investigated and suitable
modifications should be made in the case record forms.
Case Record Form (CRF) is a record of the data on each subject in a study as defined
by the protocol. The investigator must make every effort to maintain confidentiality of
information about the subjects, including the subjects identity when checking
documentation.
Data Collection & Entry: The data may be entered directly on to a computer or
collected on the CRF and then fed in. Data entry should be performed continuously
during the course of the study. It should be checked either by double entry i.e.
computerized checking or by proof reading i.e. checked manually. If the data checking
has to be done on computer, it can be combined with data entry in order to speed up
feedback. To supplement the continuous checking of each individual’s data during the
study, descriptive statistics on each important variable in the database are useful in the
detection of doubtful or unusual data.
Categories of personal information
Epidemiological research includes the use of the following types of data:
Identified: Data that allows the identification of a specific individual is referred to
as ‘identified data’. Examples of identifiers may include the individual’s name, date of
birth and address. In particularly small sets of data even information such as a postal code
may be an identifier.
Potentially identifiable (coded, re-identifiable)
Data may have identifiers removed and replaced by a code. In such cases it is possible to
use the code to re-identify the person to whom the data relates so that the process of
deidentification is reversible. In these cases the data is referred to as ‘potentially
identifiable’.
De-identified (not re-identifiable, anonymous)
The process of de-identification can be irreversible if the identifiers have been removed
permanently or if the data has been de-identified. It should be recognised that the term
‘deidentified’ is used frequently in documents other than this statement, to refer to sets of
data from which only names have been removed. Such data may remain ‘potentially
identifiable’.
OBLIGATIONS ON EPIDEMIOLOGISTS
Where epidemiological research identifies the presence of a risk to the health and safety
of particular participants or populations, the researchers should ensure that this
information is not withheld from the populations or participants affected. Feedback to
individual participants and, if necessary, family members and the community, of
information relating to health risks should be included as part of the protocol and be
required for a protocol to gain ethical approval. Epidemiologists should also take into
account the risk of publishing analyses that may stigmatize the group or groups in
question. Because epidemiologists are frequently occupied in research involving
communities and groups with high disease burdens or risks, particular attention should be
paid to ensuring that epidemiological research does not exploit the vulnerability of such
groups. Rather, epidemiological research should seek to strengthen the capacity of
vulnerable groups to reduce their vulnerability. Approval by a research ethics committee
should be contingent on a consideration of the extent to which such responsibilities have
been considered by the researcher, and where reasonably feasible, suitable actions
incorporated into the research protocol.