Documente Academic
Documente Profesional
Documente Cultură
and Demography
Epidemiological Studies
• Descriptive – analytical
• Observational – experimental
Descriptive Studies
Descriptive studies
• No hypothesis is formulated
• No analysis is performed
• Hymes, K.B., Greene, J. B., Marcus, A., et al. (1981) 'Kaposi's sarcoma in
homosexual men: A report of eight cases', Lancet 2:598-600
• Siegal FP, Lopez C, Hammer GS, Brown AE, Kornfeld SJ, Gold J, Hassett J,
Hirschman SZ, Cunningham-Rundles C, Adelsberg BR, et al.: Severe
acquired immunodeficiency in male homosexuals, manifested by chronic
perianal ulcerative herpes simplex lesions. N Engl J Med. 1981 Dec
10;305(24):1439-44.
Since prevalence is used (and not incidence), survival factors also affect
the results
The Epidemiology of Major Depressive Disorder. Results From
the National Comorbidity Survey Replication (Kessler et al, JAMA)
The Epidemiology of Major Depressive Disorder. Results From
the National Comorbidity Survey Replication (Kessler et al, JAMA)
As precisely as possible!!!
“Babies who are breast-fed have less illness than babies who are
bottle-fed.”
• Generates hypotheses
• Fast, cheap, easy to perform
• Can focus attention to special problems (region,
disease, population)
Suicide Rate
per 100,000
30
20
10
0
0 0.2 0.4 0.6 0.8 1
As non-Protestants in a province became more in the minority, these individuals,
Proportion
rather than Protestants, may have been more likely to commit suicide. Protestant
The ecological fallacy
• Statistical meaning:
The tendency of correlation coefficients to be larger
when an association is assessed at the group level
than when it is assessed at the individual level.
20
15
10
0
0 5 10 15 20 25 30
Atomistic fallacy
Examples:
• Suicide – income
Case-Control Studies
Case-control studies
Exposed
Cases
Not Exposed
Exposed
Controls
Not Exposed
Case-control Studies
• Directionality is the key determinant of whether a study
is cohort or case-control (Kramer-Boivin model).
Selection of cases:
• Incident cases
• Prevalent cases
Selection of controls
• Analysis:
• Unmatched (OR)
a
c a*d
_____ _____
OR= =
b b*c
d
Case-control studies:
Paired analysis (for one control)
Cases
Exposed Unexposed
Controls
Exposed --- R
Unexposed S ---
Strength
› Fast, relatively cheap (study population is not big)
› Can be used to study rare diseases
› Can study the effects of multiple exposures
› Helps to understand new diseases
Limitations
› No incidence rates
› No direct risk estimation is provided (no RR)
› Rare exposures can not be investigated
› Possibility of bias (recall, selection)
› Only one single disease can be studied
› Effect of matching variables covered
Case-control studies
Carcinoma No
DES 7 1
No 0 32
Case-control studies
Sample size considerations
Sample Size
Parameter
n=20 n=50 n=500
Computed
• Prospective
• Retrospective
• (Historical prospective)
(past and present exposure assessment)
Cohort studies
Comparison group
› Must be similar to the exposed cohort (confounding)
› Internal (one cohort – participants divided into exposed and non
exposed categories)
› General population rates as comparison
› Sometimes multiple cohorts needed
Cohort studies
• Analysis: Disease
+ -
Exposure
+ a b
- c d
Strengths:
• Can elucidate temporal relationship between exposure and disease
(hence, “strongest” observational design for establishing cause and effect
relationship).
Limitations:
•Current smokers
•Age started smoking
•Amount consumed currently
•Method of smoking
•Past smokers
•Same as above
•Date stopped smoking
•Outcome:
•Mortality ascertained by looking-up death certificates
•Cause of death is filled in by a physician or the coroner
•Analysis:
•Compare rates of death according to level of self-reported
smoking
Typical Questions about Smoking
• Cumulative exposure:
frequency of smoking x intensity x duration
Reasons for
nonresponse
Too ill NA 31 65 21
Refused NA 36 63 102
Not found NA 72 403 22
Other NA 369 445 362
British Doctors Study: Lung Cancer in men among
current smokers from data obtained at last questionnaire
Non smokers 10 1
Cigarettes only 140 14 (=140/10)
Pipe and/or cigars 58 5,8 (=58/10)
Mixed 82 8.2
Cigarettes only (No. per day)
1-14 78 7.8
15-24 127 12.7
≥25 251 25.1
Nested case-control study
Population
Develop Do Not
Disease Develop
Disease
Subgroup
”Cases” Selected as
“Controls”
Migrant Studies
Migrant studies
„vary environment, keep genetics constant”
With
• incidences in a group of ethnically similar individuals who have
moved to a new environment
And with
• incidences in the original population of this new environment
Cancer incidence in Japan and Hawaii
Hawaii, 1968-72
Tumor Gender Japan Japanese Caucasian
• Therapeutical
• Almost always on individual level
• Clinical trial
• Preventive
• On individual level
• Field trial
• Involving entire populations
• Community level
Study design for clinical and field trials
(Bhopal)
Factors affecting
clinical and/or field trials
• Placebo effect (tendency to report an improvement)
Randomized
Treatment A Treatment B
Group 2
Group 1
Group 2
Group 1
Group 2 Group 1
Crossover trial
• CARET
• ATBC
On beta-carotene supplementation
Retrospective studies: Dietary carotenoids and lung cancer risk
mg/day
35
30
25
20
15
10
5
0
Men Women Supplement
Factors affecting the antioxidant/prooxidant
properties of -carotene
• Concentration
• Presence of certain prooxidant compounds
• Oxigene-concentration
• pO2 below 50 Hgmm (peak at 15 Hgmm) - antioxidant
• Between 50-100 Hgmm looses its antioxidant activity
• Above 100 Hgmm prooxidant
• Presence of other antioxidants (e. g. Vitamin E)
Evidence Based Medicine:
Evidence-pyramide