Sie sind auf Seite 1von 45

2.

ANALYTICAL STUDY

COMPARATIVE or ANALYTICAL STUDY


An ANALYTICAL STUDY attempts to establish causes or risk factors for certain problems. This is done by comparing two or more groups, some of which have or develop the problem and some of which have not.

Analytic Epidemiology
Second major type of epidemiological studies
Subject of interest is individual within the population The objective is to test hypothesis The study determines whether or not a statistical association exists between a disease and suspected factor Strength of association, if it exists

CASE CONTROL STUDY


In a CASE-CONTROL STUDY, the investigator compares one group among whom a problem is (e.g., malnutrition) with another group, called a control or comparison group, where the problem is absent to find out what factors have contributed to the problem.

CASE CONTROL STUDY


Often called retrospective study

First approach to test causal hypothesis

Properties
Both exposed and outcome (Disease) have occurred before the start of study The study proceeds backward from effect to cause It uses a control or comparison group to support or refute an inference

Case Control VS Cohort Study


CASE CONTROL STUDY

Factors
Present Absent

Disease
Present (Cases) Absent (Controls)

Risk Factors
Exposed Unexposed

COHORT STUDY Disease


Present on Followup Absent on Followup

TIME LINE

2 x 2 Contingency Table for Cases and Controls


Cases Exposure + Exposure Total Cases Controls

A
C

B
D

A+C

B+D
Exposure Among Controls
B/(B+D)

Exposure Among Cases


A/(A+C)

Method
Selection of cases and Controls Matching

Measurement of Exposure
Analysis and Interpretation
Exposure Rates Estimation of Risk (Relative Risk & Odds Ratio)

Selection of Case and Controls


Selection of Case Cases can be selected from a variety of sources including hospital patients, private clinics or general population. Many communities maintain registries of patients with certain disease, such as cancer and these registries serve as valuable sources of cases.

The entire case series or a random sample of it is selected for study. General population: In a population-based case control study, all cases of the study disease occurring within a defined geographic area during a specified period of time are ascertained, often through a survey, a disease registry or hospital network.

Case Definition Appropriate case definition should be made. Diagnostic criteria for case selection should be clearly specified. Eligibility Criteria It is considered that for better study results incident cases (newly diagnosed cases) are more eligible than old cases in an advanced stage of disease.

Selection of controls
Controls must be free from the disease under study. Sources of Control Hospital controls Relatives Neighborhood General population

Number of Controls If the study group is large one control may be identified for each case. For small study groups (<50) as many as 2,3 or even 4 controls may be selected for each case.

Matching
Process of selecting controls so that they are similar to the case in characteristic such as age, race, sex, socioeconomic status and occupation.

Matching is done to ensure comparability between cases and controls. In other words matching is done to eliminate confounding factors.

Confounding Factor This factor although associated with exposure under investigation is an independent risk factor for the disease in itself.
Alcohol Oesophageal Cancer

Smoking

Examples: 1. In the study of role of alcohol in the aetiology of oesophageal cancer, smoking is a confounding factor because it is associated with alcohol consumption and smoking is an independent risk factor for oesophageal cancer. 2. Age could also be a confounding variable. If women taking oral contraceptives were younger than those in the comparison group, they would be at lower risk of breast cancer

Type of Matching
Matching may be of two types: Group matching Individual matching

Group Matching (frequency matching):


Controls are selected in such a manner that the proportion of controls with a certain characteristic is identical to the proportion of cases with the same characteristic. Thus, if 25% of the cases are married, the controls will be selected so that 25% of that group is also married.

Individual Matching (matched pairs):


For each case selected for the study, a control is selected who is similar to the case in terms of the specific variable or variables of concern, for example, if the first case enrolled in a study is a 45-year-old white female, control also will be a 45-year-old white female.

Measuring Exposure
Interviews on exposure is measured by using questionnaires, past records (hospital, employment etc). This should be done in a precise manner for both cases and controls.

Estimating Risk
Odd ratio: It is the ratio of the odds that a case was exposed to the risk factor to the odds that a control was exposed to the risk factor. Odds ratio is a cross product ratio. Exposed with disease x unexposed without disease Formula =
Exposed without disease x unexposed with disease Disease Yes No

Exposure

+ a - C

b d

ad = bc

Significance: OR provides a reasonably good estimate or relative risk provided that the: Incidence of disease is low (as in chronic disease). Cases and controls are representative. the odds ratio is similar to relative risk as both demonstrate the strength of the association between the risk factors and the disease (but in different ways). Odds ratio is also referred at time as estimated relative risk.

Inference
OR = 1: Risk factor not related to the disease. Risk factor is a actually protective factor against the disease. Risk factor positively associated with the disease.

OR < 1:
OR > 1:

Relative Risk (RR)


Ratio of incidence of the disease (or death) among

exposed and the incidence among non-exposed.


It is a direct measure (or index) of the strength of the

association between suspected cause and effect

Odds Ratio (OR)


Measure of the strength of the association between risk factor and outcome. The derivation of the Odds Ratio is based on three assumptions: - the disease being investigated must be relatively rare - the cases must be representative of those with the disease - the controls must be representative of those without disease

Analysis
Relative risk (RR) Ie RR = ----Io Attributable risk (AR) Ie-Io AR = -------- X100 Ie

Whereas
Ie : Incidence among exposed Io : Incidence among non exposed

Role of relative risk in case-control studies


Incidence rates cannot be obtained from a typical case control study from which relative risk can be calculated since there is no appropriate denominator or population at risk. In general relative risk can be exactly determined only from a cohort study.

Oral contraceptives and Thromboembolic disease


Period: 1968 1969 Cases: Women admitted in hospitals with venous thrombosis or pulmonary embolism without medical causes. Controls: Women admitted to the same hospital with other disease matched for age, marital status and party. Conclusion: Users of oral contraceptives were 6 times likely as nonusers to develop thromboembolic disease.

Thalidomide Tragedy
Study Period: 1958 1961 Cases: 46 mothers who delivered deformed babies. Controls: 300 mothers who delivered normal babies. conclusion: 41 out of 46 mothers had thalidomide during their early pregnancy.

Bias in epidemiological studies


These are systematic errors occurring in epidemiological studies. About 35 different types of biases exists. The important bias are: Selection bias Information bias (Recall bias) Interviewers bias Confounding bias Survival bias

1) 2) 3) 4) 5)

Selection bias:
This is due to systematic difference between the characteristics of the people selected for the study and the characteristics of those who are not.
For example In a clinical trial for evaluating new drugs for treatment of myocardial infarction, if older people are excluded from the group getting the new drugs, this will lead to a selection bias (because older people have poorer prognosis and they are intentionally not included in study group).The bias

Information bias (recall bias)


Individual with a disease tend to report past events more accurately than those without a disease leading to recall bias.

Interviewers bias
If the interviewer tends to probe for answers more while interviewing cases as compared to interviewing controls, this will lead to information bias.

Confounding bias
While studying association between exposure and outcome (disease) confounding can occur when another exposure exists in the study population and is associated with both the disease and the exposure being studied. For example, if an association is found to exist between coffee drinking and coronary heart disease, then this association may not be a true association due to confounding.

The confounding factor here could be cigarette smoking, people who drink coffee are more likely to smoke than people who do not drink coffee. It is well known that smoking is associated with CHD. It is thus possible that the relationship between coffee and CHD is actually the reflection of strong association between smoking and CHD. Smoking here confounds the observed association between coffee drinking and CHD.

Berkson bias This bias occurs due to different admission rates in hospitals. Each hospital will have an independent admission policy and therefore certain disease may be given more priority in that hospital. This type of bias is commonly taken into account in case-control studies.

Advantages of case control study


Relatively easy to carry out Rapid and inexpensive Particularly suitable to investigate rare diseases No risk to subject Reveals the study of several different etiological factors Risk factors can be identified No follow up in the future

Disadvantages of case control study


Problems of bias Selection of appropriate case control group may be difficult Cannot measure incidence, only relative risk is measured

Practical Application
case control studies are used for investigating the causes of disease particularly rare diseases. Two groups of people are studied, one group who have the disease under investigation and the other who are free from disease. The occurrence of the possible cause of the disease is compared between the cases and controls.

As a registrar in the paediatrics department of a teaching hospital you observe that during the past one year there has been an unusually high number of children who have been born with birth defect. On preliminary investigation it is seen that the mothers of these children had taken a drug which had recently been introduced in the market and was supposed to prevent nausea int eh first trimester

a) How would you test your hypothesis (hunch) in order to implicate the drug as a probable cause of the limb defects? b) How would you select your controls? c) Would matching your controls be necessary in this study?

Randomized controlled trials


Investigator controls the predictor variable (intervention or treatment) Major advantage over observational studies is ability to demonstrate causality Randomization controls unmeasured confounding Only for mature research questions

Das könnte Ihnen auch gefallen