Statistics

EPIDEMILOGICAL STUDY DESIGNS

Dr Faris Al-Lami MB,ChB MSc PhD FFPH

Definition of Epidemiology

"Epidemiology is the study of the distribution and determinants of health related states or events in specified populations, and the application of this study to control of health problems.”- Last’s Dictionary of Epidemiology

USES OF EPIDEMIOLOGIC RESEARCH METHODS

1. Measure the disease frequency and the pattern of disease occurrence 2. Identification of disease determinants 3. Evaluate validity of diagnostic screening tests 4. Determining the natural history of the disease and clinical course 5. Determining the prognostic factors 6. Testing new treatments

Epidemiological studies can be classified into:

I. Descriptive studies which can be classifies into: 1. Studies conducted on population (Correlation studies) 2. Studies conducted on individuals: a- Case report b- Case series c- Cross sectional surveys

Epidemiological studies can be classified into:

II. Analytic studies which can be classified into: 1. Observational analytic studies: A-Case control studies B-Cohort studies 2. Interventional studies (Clinical trials)

Epidemiological Studies

Descriptive studies
Analytic studies
Population
Individuals
Correlation Studies
-Case Report -Case series -Cross Section Studies
Observation Studies
Intervention studies
-Case Control Studies -Cohort studies
Experimental studies (clinical trials)

The Five W’s of Journalism / Epidemiology What Who Where When Why / How

Diagnosis (the disease or condition being investigated) Person (who is getting the disease, who is at risk) Place (residence, work, school, etc.) Time (date and/or time of onset of disease, time of exposure) Cause, mode of transmission, risk factors

The Five W’s of Journalism / Epidemiology What Who Where When Why / How

Clinical Person Place Time Cause, mode of transmission, risk factors
Descriptive Epidemiology (Distribution)
Analytic Epidemiology (Determinants)

DESCRIPTIVE STUDIES

Advantages: They use already available data They are less expensive and less time consuming They describe the pattern of disease occurrence They formulate research hypothesis

DESCRIPTIVE STUDIES

These studies are essential for: 1. Public health administrators: To determine which population or subgroup of the population are most or least affected. To decide on efficient allocation of resources and preventive programs. 2. Epidemiologist: To identify risk factors

I. Correlation or Ecological Studies

Based on aggregate measures of exposure and outcome from several populations. The population is the unit of observation available for study.


Ecological Study of Homicide Rate by Percentage of Households with a Gun in 13 Countries, 1990
*per million person-years

I. Correlation Studies

Advantages:Describe the disease in the entire population in relation to factor of interest.Use the Correlation Coefficient (r) to measure the association between the two variables of interest.Easy to do, inexpensive, and can be conducted quicklyRepresent the first step in searching for exposure disease relationship (Generate hypotheses) Use available data (“administrative” or other aggregate data)

Correlational or Ecological Studies

Exposures: - What percent of a population smokes? - What percent of 1-year old children are vaccinated against measles? - What percent of a population has piped water?
Outcome: - What percentage of a population died from MI? - What percentage of children had measles last year? - What percentage of population had episodes of diarrhea?

Correlational or Ecological Studies

Disadvantages Correlation data represent average exposure level rather than actual individual values. Data on exposure and data on outcome are collected independently No assurance that persons with exposure (risk factor) of interest are the same ones with the outcome (disease) of interest

Disadvantages

Inability to link exposure with disease in a particular individual. Association at the aggregate level may not reflect association at the individual level - the ecologic fallacy Unable to adjust for potential confounding factors.

II. Case report and case series

The individual is the unit of observation. They describe the experience of a single patient or a small number of patients with a similar diagnosis reflecting unusual features of a disease. They help in: formulation of a hypothesis on etiological association represent the first clues in identification of new disease or epidemic.

Case reports

Clinical case with “unusual” clinical picture (e.g. 40 y.o. female, pulmonary embolism and no history of CVD, recent use of OC)

Case series

First case report may stimulate compilation of additional case reports….a case series(e.g. occurrence of Pneumocystis carinii among a group of young, homosexual men with no history of immune deficiency)

Case reports or Case series

Advantages: Use available clinical data Detailed individual data Suggest need for investigation (hypothesis generation) Disadvantages: May reflect experience of one person or one clinician No explicit comparison group

III. Cross sectional surveys

Collection of data on several individuals at “one point” in time. Determines prevalence at a point in timeTherefore, CS is a prevalence studyThe exposure and disease status are assessed simultaneously among individuals in a well defined population. Snapshot in timee.g. - cholesterol measurement and ECG measured at same time

Cross sectional studies can be used to formulate a hypothesis but not to test it EXCEPT if the exposure is present since birth as blood group, race, eye colour,

Cross-sectional Surveys

Examples: national census, community survey, survey of a particular group of persons (e.g., occupational group, school children)

III. Cross sectional surveys

Advantages: Provide information on frequency and characteristics of the disease. Determine prevalence of disease or any health phenomenon. Standardized data collection tool.


Advantages:
Able to focus data collection in specific locations or specific groups of persons. Can make comparisons among study participants. Relatively quick to do. Can be repeated to get data on trends.

III. Cross sectional surveys

Limitations:Inability to determine the temporal relationship between exposure and disease.May be biased by lack of participationReflects prevalent, not incident cases and thus results may be related to duration of disease, or survival with disease (e.g., “healthy worker effect”)

Job A (hazardous) 80 healthy 80 well 100 workers 20 ill 10 ill

Job B (non- hazardous) 95 healthy 95 well 100 workers 5 ill 15 ill
Prevalence ill job A: 20/100 = 20% ? Prevalence ill job B: 5/100 = 5% ? Prevalence ratio: 4 ?
Cross-sectional Studies
10
Time

How to conduct a Cross-Sectional Study

1. Clear definition of study question. i.e.: Exposure and outcome 2. Identification of study population: Small population All are included  Large population Sampling

How to conduct a Cross-Sectional Study

3. Data collection: Exposure and outcome 4. Data analysis: Subdivision of the population to suspected factor and compare the prevalence of the outcome in each subgroup 5. Interpretation: is there a relation between exposure and outcome?


Design of a C-S Study