What is Epidemiology? Briefly examine various categories of Observational studies.

 The term ‘epidemiology’ is derived from three Greek words namely ‘epi’(upon), ‘demos’(people) and ‘logos’(study) collectively meaning ‘study upon people’. This branch of science studies all aspects of human health and diseases. Epidemiology is defined as “the study of the distribution and determinants of health- related states or events in specified populations, and the application of this study to the prevention and control of health problems.”

Observational Studies

In observational studies, frequency and distribution of diseases or deaths are reported by time (year/month/week/day/hour/season), place (country/urban-rural/ institutions/hospitals/old age homes/ schools) and demographic characteristics (age/sex/income/education/occupation/marital status, religion/caste). Observational studies are categorised into two types – (i) Descriptive and (ii) Analytical studies. In descriptive studies, only description about the disease (case reports/case series) is made whereas in analytical studies (ecological/case-control/ cross-sectional/cohort) relationship of variables (causative factors) with diseases is described.

 

1) Descriptive Studies

Descriptive studies are again of two types, case reports and case series.

i) Case Reports

In case reports, cases with unusual symptoms, signs and characteristics or death observed during clinical practice are reported by the clinician’s presentations which are helpful to define new clinical disease/entity. These case reports are useful in clinical practice, formulating hypothesis and explore in epidemiological studies. Example: coagulopathy in patient with renal failure.

ii) Case Series

When new clinical entities/new cases or deaths with common characteristics, symptoms or signs are compiled by single or group of clinicians they can be called case series. They are useful for definition of new cases, to understand the spectrum of symptoms and signs, when followed till the death of patients which are useful to investigate the natural history of disease. The data are usually collected from clinicians and sometimes from populations in case of sudden deaths within a defined geographical area. Data of case series can be used to know the distribution of disease by place, time, religion, ethnicity, season and socio-economic status. Acquired immune deficiency syndrome is defined as new disease after publication of case series of young men contracted with pneumocystis carinii pneumonia and Kaposi’s sarcoma. Case series data can be used to formulate hypothesis, easy to collect, cost effective and quickly available. Case-series data cannot be used to calculate rates of disease as no denominator is available, involve no comparison group, suffer from sampling variation and recruit only selective cases. Example case series on symmetrical acrokeratoderma (dermatosis).

 

2) Analytical Studies

Analytical studies are of four types :

i) Ecological Studies

In this type of studies, association between disease/outcome frequency and the level of exposure in groups of within or between populations is studied. Population not the individual is the unit in this kind of study. Grouping can be done based on the place (birthplace/residence/factory/school), socio-economic status, time or by mixing place and time. Ecological studies are used for generation of hypothesis. Data from public/private sources, registries/death certifying organizations and earlier surveys can be used. For example, in this type of study, investigating the incidence of cancer in different countries can obtain the details on age distribution and disease status from census data and tumour registries. Example: spatial spread of leprosy in India.

ii) Case-control Studies

These studies investigate the aetiology of disease, suitable for studying rare and longer duration of (chronic) diseases, cost effective, require less number of subjects, easy to perform, no risk is done to subjects, multiple risk factors can be studied at the same time, no dropout of subjects is observed, has minimal ethical problems and can be completed within short duration. The unit of study is individual. Newly diagnosed cases are compared with subjects without disease. Exposure to potential risk factors in both cases and controls is evaluated by examining case sheets/enquiry of patients or patient relatives/controls or by performing biochemical tests. These studies are called retrospective (as the study deals backwards from outcome/disease to cause) and prospective (if the data collection is still in progress).

 

Cases are recruited from hospitals/patient registries/cross sectional study/caseseries/cohort study. Controls are drawn from same geographic area/spouses/ friends, from same office/factory/institute or patients diagnosed with other disease from same hospital. Cases and control can be matched for age, sex, and ethnicity, social class (income, education and occupation) to reduce selection bias. The association between exposure (causative agent/risk factor) and disease/outcome is evaluated by Odds ratio. Demerits of case-control are difficulty in finding suitable controls, subjects may not be representative of population, prevalence/ incidence or attributable risk cannot be estimated, efficacy of therapeutics cannot be evaluated, not possible to distinguish between causative or accompanying factors, suffers from confounding(due not mismatching of subjects), recall (cases more likely recall the presence of events), selection (subjects not recruited as per standard criteria), Berkesonian (recruitment of subjects from sub population than general population) and interviewer bias. Example: Utility of anthropometric traits and indices in case-control study.

 

iii) Cross-sectional Studies

In these studies, both exposure and outcome (disease) are investigated at the same time. No temporal associations between exposure (risk factors) and outcome can be explained. The unit of the study is individual. These studies are useful for investigating chronic diseases and fixed exposures such age, gender, ethnicity and genotype, to study multiple risk factors simultaneously. These studies are easy to conduct, give inputs on burden of disease which can be used for planning health infrastructure, allocating resources and manpower. It is inexpensive and can be completed within a short period of time. If cross sectional study is repeated on the same population it can serve as cohort study and if repeated on independent sample, it is useful to investigate the trends of the disease. For variable exposure, data on past and present exposures are recorded.

 

The target population is studied using representative population and the results are extrapolated to this population. These studies are also called prevalence studies. If prevalence is standardized using the data of standard populations, the prevalence can be compared with other populations. Both disease and determinants can be studied in this type of studies. To avoid sample bias, random sampling techniques are used such as simple, systematic, clustered, stratified, multistage and mixed. The denominator is usually the population at risk or total population studied. Prevalence is presented a per cent or per 1000 subjects. Prevalence studies are of three types depending on the time involved. They are point, period and lifetime prevalence.

 

Prevalence studies are not suitable for studying the natural history of disease and to estimate the incidence. Subjects deceased or with severe disease are missed out in these studies. It is not possible to distinguish whether risk factor or exposure precedes the outcome/disease or exposure is resulted from the outcome. If investigator fail to gain the confidence of subjects which results in high nonresponse rate resulting leading to selection bias. This type of study design is not suitable for rare diseases. A statistical technique called Logistic regression analysis can be used to find the association between risk factors and disease. Example: Prevalence of coronary artery disease and coronary risk factors in Tirupati urban population.

 

iv) Cohort Studies

These studies are called incidence/longitudinal studies. Cohort means group of population. Groups can be formed based on the date of birth (birth cohort), date of marriage (marriage cohort), decade (decade cohort), occupation (doctors/ lawyers/engineers/teachers), city population (Example: Delhi) etc. Subjects of cohorts have common characteristics/experience/condition. A group is assigned for the study, and exposed and non-exposed cohorts within the same group are identified and followed for particular period. If exposure is rare, this cohort is compared with external cohort matching all characteristics except exposure. Same cohort can be divided into subgroups based on level of exposure and outcome. Diagnostic criteria for outcome of interest are decided at the beginning of the study. Baseline data is collected from cross sectional studies, census and birth registries. Data on exposure are collected by conducting interviews, contacting subjects on mobile/through e-mail, examination of case sheets, conducting of diagnostic tests and environmental surveys such as air or water quality. Subjects with disease are excluded from this study. Both exposed and non-exposed cohort subjects are evaluated periodically on clinical status, performing diagnostic tests, reviewing cash sheets and visiting the subjects for examining the end points (outcome/disease/death) of study. Presence of outcome/disease/death is compared between exposed and non-exposed cohorts. Incidence rates, relative risk (measure of evaluating the strength of association between exposure and outcome), attributable risk (what extent disease is due to exposure) and population attributable risk (suggest to what extent disease is reduced if the exposure is eliminated) are determined and compared between both cohorts. Cohort studies are of three types namely prospective (outcome occur after initiation of the study), retrospective (outcome occurred before the initiation of the study) and mixed (outcome occurred before the initiation of study which is further assessed prospectively). If the newly identified cases in the cohort study if compared with control of the same cohort, it is called nested case-control study. Example for Prospective cohort study of overweight and obesity in rural population of West  Bengal, India and for retrospective cohort study example is maternal and neonatal outcome of gestational diabetes in the subjects of Kerala; Mixed cohort study example is retrospective and prospective cohort study on HIV sero status and incident pneumonia

Subcribe on Youtube - IGNOU SERVICE

For PDF copy of Solved Assignment

WhatsApp Us - 9113311883(Paid)