1. epidemiology, rct, screening, biostats

Epidemiology term given by

John M Last

Epidemiology define

FDDC

study of Frequency Distribution Determinants & application of knowledge for disease PREVENTION & CONTROL

Disease distribution means?

Time , place , person of distribution of disease

TPP

In rate is N part of D

Yes

Rate = time

1k, 10k or 1lac

In ratio is N part of D

No

Ratio= comparison

In Proportion, is N part of D

Yes

Part of whole = proportion

Always 100

Incidence formula

New cases / population at risk x 1000

Incidence is RATE, COHORT

Prevalence formula

Existing cases / total population x 100

Prevalence is PROPORTION, CROSS SECTIONAL

Descriptive study (buzz word)

Distribution

Analytical study buzz word

Determinant

What is field trials

Vaccine trials

Done on: HEALTHY PEOPLE

What is community trials

Preventive trials

Formulation of hypothesis is ?

Descriptive epidemiology

Test of hypothesis is

Analytical epidemiology

Confirmation of hypothesis is

experimental & interventional epidemiology

Ecological study is for

Population (correlation, geographic , aggregate)

New disease with unknown cause is reported, which study would u do

1st : Descriptive study

Cohort is which type of epidemiological study?

Analytical

Cohort = analysis

“Matching” is the buzz word for ?

Case control

Rare exposure : which study?

Cohort

Rare diseases : which study?

Case control

Study done at same point of time

Cross sectional

Exposed vs non exposed study

Cohort

Diseases vs non diseased study

Case control

“Follow up” is the buzz word for

COHORT

Study that analyses cause to effect

Cohort

Largest prospective study

Framingham heart study

3 generations of study

Hill’s criteria is for which study?

Cohort

Advantages of cohort

TMID

TEMPORALITY (cause proceeds effect)

Multiple outcomes (CVD, ca, etc)

Incidence calculation

Direct RR

Rare exposures

Minimal BIAS (eg , recall bias)

Hawthorne bias (cohort)

Subjects modify behaviour

Eg: quit smoking in between

Hawww chod dia

Attrition bias (cohort)

Loss to follow up

Ambispective cohort

Both retrospective and prospective study

Advantages of retrospective cohort

Save time and money

No attrition bias ( loss to follow up)

No Hawthorne bias (changed behaviour)

What is longitudinal study

Outcomes are looked at fixed time intervals (eg: every6months)

Ensuing meaning

In coming up…

Ensuing years upcoming yrs

RR, attributable risk, PAR

These belong to which study

Cohort

Relative risk formula

incidence(exposed)/ incidence (non exposed)

RR = a/(a+b) ÷ c/(c+d)

RR >1, =1, <1

>1 - positive association (risk)

=1 - no effect

<1 - negative association (protective)

attributable risk meaning

AR= 75%

75% of lung cancers are attributed to smoking

EPIDEMIOLOGIST : use it

Population attributable risk (PAR)

“If we REMOVE exposure → how much disease reduces?”

PAR=90%

If smoking is stopped , 90% reduction in annual incidence of lung cancer cases in the population

Used by: POLICY MAKERS

community physicians / epidemiologists use which risk,

Attributable risk

Policy makers use which risk ?

PAR

Incidence of disease in exposed vs incidence of disease in non exposed =

Relative risk (cohort study)

Buzz words for CASE CONTROL STUDY

matching

Opposite of cohort

Trohoc = case control

How many controls can be taken for 1 case in case control study

Max 4

1:1 is also ok

Advantage of matching

Eliminates all the confounders of the study

Example of case control study

Dolls & Hills study

Bias in case control study

Recall bias

Measure of strength of association of case control study?

ODDS RATIO/ CROSS-PRODUCT RATIO

Formula = ad/bc

Odds ratio and RR relationship

OR provides an estimate of RR

Nested case control

Chota case control

Bada cohort

Chota study inside bada cohort

Nested case control features

Good for RARE INVESTIGATIONS

TEMPORALITY established (cause proceeds effect)

Saves MONEY and TIME

Study of biological precursors of a disease

Nested case control study avoids which problem as compared to a case control study?

TEMPORALITY

Cross sectional study buzz words

Snapshot (same time)

PREVALENCE : used for descriptive purposes

Biggest disadvantage of CROSS SECTIONAL STUDY

cannot establish TEMPORALITY

(Snap se kaise pata chalega ki cause kya h snap ka ?)

Screening test / health camp is which type of study

Cross sectional ( same time p snapshot)

Ecological study aka?

Correlation study

Geographical study

Population study

Aggregate study

Buzz words for ecological study

Country wise

State wise

Unit

2° data

Which type of study is used for nutritional services ?

Ecological (nutrition — sabka —- population)

Biggest problem in ecological studies

Ecological fallacy ( cannot apply population data on individuals)

We generalise the outcomes of a population on individuals which might not be correct !

High meat consumption —> colorectal cancer higher in those countries

But people having colorectal cancer may not be eating meat

Which graph is used to express ecological studies

Scatter diagram

Concept of hospice care is to give supportive care to

Final phase of terminally ill patient

Web of causation of disease helps to?

Suggest ways to interrupt the risk of transmission

Epidemiological ratio

DALY (CDs) / DALY (NCDs)

Summary of epidemiological studies . Is called ? Is it qualitative or quantitative study?

Systematic review

Qualitative only

Does NOT include quantitative synthesis of data of all studies available

Summary of a study + Mathematical analysis (quantitative) = ?

Meta analysis

Meta = Math

All meta analysis is based on systematic review but .

But all SR ≠ meta-analysis

Steps of systematic review

DSS ESI

Define ques

Search study

Select study

Extract data

Synthesise data

Interpret

funnel plot

Shows combined effect of SR & MA

Checks for PUBLICATION BIAS **

File drawer effect (funnel plot)

Dots NOT uniformly distributed = publication bias +

No file drawer effect (funnel plot)

Dots uniformly scattered = no publication bias (study publish na hona)

Funnel straight = no bias

Tilted = bias

Other name of forest plot

Blobbogram

ONLY made for META ANALYSIS (MA)

Only MA FOREST m BUBBLEGUM khila skti h

How to interpret forest plot

If CI crosses 1 → ❌ Not significant

👉 If CI does NOT cross 1 → ✅ Significant

>1 = risk factor

<1 = protective factor

What does DIAMOND mean in forest plot

Overall estimate / pool estimate

Diamond = final ans

Reporting guideline of : case report

CARE

Care = case report

Reporting guideline of : RCT

CONSORT

Consort = RCT

Reporting guidelines of : observational studies (case control , cohort , cross sectional

STROBE

Strobe = study observational

Reporting guidelines of META analysis

MOOSE

MOST IMPORTANT!!

Reporting guideline of : systematic review + MA

PRISMA

PRISMA = pooled studies

Father of evidence based medicine

David sackett

Gold std of evidence based medicine

Meta analysis

Founder of evidence based medicine

Gordon guyat

Founder guyat

Father sackett

Hill’s criteria gold std study

Meta Analysis

Most imp criteria in hills criteria

TEMPORALITY

(Cause to effect)

Least imp criteria in hills criteria

Specificity (one exposure , one outcome )

(Smoking only leads to lung cancer)

Give examples of strength of association?

RR= cohort

OR= case control

Case series

Collection of case report

Acupuncture in children w chronic headache study

Hills criteria is used to

Establish CASUALITY ( proving that this factor causes the disease , not just associated with it)

Random in randomization = ?

Equal and known chance

Intention to treat analysis is done in?

RCT

Major purpose of random assignment in a clinical trial ?

Ensure that study grps are COMPARABLE on baseline characteristics (eliminates selection bias)

Double blind = ?

Both observer & observed grp is blind

Best study design to establish CAUSALITY (cause→effect)

Non concurrent cohort ( retrospective cohort)

Co founder = ?

Linked to both exposure and disease

Eg: LIGHTER (linked to both smoking and lung cancer ) but independent risk factor h

Effect modifier = ?

If a person is exercising, effect of smoking on lung cancer reduces

A variable that (positively or negatively) modifies the observed effect of a risk factor on a disease status

Techniques to eliminate confounders

RANDOM = removes all

MATCHING = removes small

Which technique removes known and unknown confounders

Randomisation (best)

What is restriction in observational studies ?

Limit the study to one category . Eg: only males

matching

Case control

Eliminates ONLY KNOWN confounders