Transcript Randomized Controlled Trial

Siroli Lily: State Flower of Manipur
Crosssectional
Study
Subodh S Gupta
MGIMS, Sewagram
Simplest research questions
What is the total population of Imphal?
► What proportion of married men in
Imphal help their wives in kitchen.
► What proportion of patients attending
OPD in Imphal hospital and research
center come for psychiatric disorders?
► What is the prevalence of underweight
among under-five children in Ukhrul?
► What proportion of residents of Imphal
above 30 years of age do exercise
regularly?
►
More research questions
Do married men younger than 40 years
help their wives in household chores
more often than those above 40 years?
► Is the prevalence of hypertension
higher in those who exercise regularly
than those who do not exercise
regularly?
►
Types of
study
design
Grimes and Schulz. Lancet 2002
Snapshot observation Vs longitudinal
observation
Cross-sectional studies
► Descriptive,
or
► Analytical, or
► Both
Gynecomastia in a drug company
► Puerto
Rico pharmaceutical company:
Survey showed that employees had
gynecomastia
► OC pills; oestrogen dust might be the cause
► Dust control measures; epidemic
disappeared
Harrington et al. Arch Environ Health 1978; 33: 12-15
Demographic surveys:
a type of cross-sectional studies
►
►
►
►
National Family Health Survey
District level Health Survey
NNMB Survey
Sentinel surveillance for HIV
HIV prevalence in India
by district, 2005
Multiple cross-sections help in
giving a whole picture?
Estimated adult HIV prevalence &
number of PLHA, India, 2004-09
HIV estimation 2010
Uses of cross-sectional studies
Public Health
► Community diagnosis




Health status
Determinants of health & disease
Association between variables
Identification of groups requiring special care
► Surveillance
► Evaluation
of community’s health care
(coverage evaluation)
Uses of cross-sectional studies
► Individual
& family care
► Studies on diagnostic test
► Studies of growth & development
Cross-Sectional Studies
Advantages
► Cheap
and quick studies.
► Data is frequently available through current
records or statistics.
► Ideal for generating new hypothesis
► Generalizable results if population based
sample
► Study multiple outcomes and exposures
► Can measure prevalence
► Hypothesis generating for causal links
► Serial surveys
Cross-Sectional Studies
Disadvantages
The importance of the relationship between
the cause and the effect cannot be
determined.
► Temporal weakness:
 Cannot determine if cause preceded the effect
or the effect was responsible for the cause.
 The rules of contributory cause cannot be
fulfilled.
► Impractical for rare diseases if pop based sample
Prone to bias (selection, measurement)
Sampling methods
► Probability




sampling
Simple random sampling
Systematic sampling
Stratified random sampling
Cluster sampling
► Non-probability
sampling
 Consecutive sampling
 Convenience sampling
 Purposive (Judgmental) sampling
Specifications & Sampling
Accessible
population
Target
population
Target
population:
Intended
Sample
Clearly defined
clinical &
demographic
characteristics well
suited to the
research question
Example: Hypertension among adults (aged 18
years and above)
Specifications & Sampling
Accessible
population:
Accessible
population
Target
population
Intended
Sample
Specify temporal
and
geographic
characteristics
representative
of target
populations
and easy
to study
Example: Hypertension among adults (aged 18
years and above in the field practice area of MGIMS)
Specifications & Sampling
Accessible
population
Target
population
Intended
population:
Intended
Sample
Design an approach
to select a sample
representative of
accessible
population & easy to
do
Example: Hypertension among adults (aged 18
years and above in the field practice area of MGIMS)
Precision & Accuracy
Good precision
Poor precision
Good precision
Poor precision
Poor accuracy
Good accuracy
Good accuracy
Poor accuracy
Confounding
► Example
Criteria for confounding
1.
2.
3.
The confounder must be associated with
the exposure
The confounder must be associated with
the disease, independent of the exposure
The confounder must not be part of the
causal pathway connecting the exposure to
the disease.
Example
► Crude
analysis
Criteria 1
► Stratified
analysis
Example:
► Criteria
1: The confounder must be
associated with the exposure
Example:
► Criteria
2: The confounder must be
associated with the disease, independent of
the exposure
Bias in cross-sectional studies
Selection Bias (eg, NSSP study)
Is study population representative of target
population?
Is there systematic increase or decrease of RF?
Measurement Bias
Outcome
► Misclassified (dead, misdiagnosed, undiagnosed)
► Length-biased sampling
 Cases overrepresented if illness has long duration and are
underrepresented if short duration.(Prev = k x I x duration)
Risk Factor
► Recall bias
► Prevalence-incidence bias
 RF affects disease duration not incidence eg, HLA-A2
Analysis
►Analysis
plan
 Depending on objectives of the study
 Dummy tables
Analysis- Descriptive CS study
► Objective:
 To describe the disease in time, place and
person
 To generate hypothesis
► Analysis





Means & SD
Median & percentile
Proportions – Prevalence
Ratios
Age, sex or other group specific analysis
Analysis – Analytical CS study
► Objective:
 Is there any association?
 What is the strength of association?
► Analysis:
 Is there any association?
 What is the strength of association?
►Correlations
►Regression coefficients
►Differences between
►Odds ratio
►Risk ratio
►Risk difference
mean
Other analysis
► Stratified
analysis
► Logistic regression