Cohort studies

PROGNOSIS
Cohort studies
Cohort studies
Useful for
z Assessing prognosis, natural history of
disease
z Assessing risk factors
z Testing aetiological hypotheses
Distressed 3100g male
newborn
z
At delivery
z
z
z
z
z
Limp, cyanotic, no spontaneous respiratory
effort, heart rate 50 beats/min.
Suction - no grimace, cough or sneeze
Vigorous resuscitation
At 1 minute: Apgar score = 1
At 5 minutes: Apgar score = 2, 110
beats/min.
z
At 10 minutes: Apgar score = 3
Score
The values for each of the five categories are added to yield a result
from 0 to 10
Sign
0
1
2
Heart rate
(beats/min)
Absent
<100
≥100
Respiration
Absent
Slow, irregular
Regular, crying
Muscle tone
Limp
Slow flexion
Active motor
Blue, pale
Body pink,
extremities
blue
Completely pink
None
Grimace
Cough, sneeze
Colour
Reflex response
to catheter in
nostril
APGAR SCORE FOR EVALUATION OF NEONATAL ASPHYXIA
Adapted from Apgar V, James LS. Am J Dis Child 1962; 104:419
Distressed 3100g male
newborn
z
At delivery
z
z
z
z
z
Limp, cyanotic, no spontaneous respiratory
effort, heart rate 50 beats/min.
Suction - no grimace, cough or sneeze
Vigorous resuscitation
At 1 minute: Apgar score = 1
At 5 minutes: Apgar score = 2, 110
beats/min.
z
At 10 minutes: Apgar score = 3
z
PROGNOSIS??
Objectives
z
Cohort study
z
z
z
describe structure and analysis
interpret and calculate relative and absolute
measures of effect
identify potential for bias
RCT / Cohort
DISEASE
EXPOSED GROUP
(eg treatment)
NO DISEASE
DISEASE
NON EXPOSED
GROUP
(eg no treatment)
NO
DISEASE
START
Apgar scores as predictors of chronic
neurologic disabilities
z
Nelson KB, Ellenberg JH, Pediatrics, 1981; 68:36
z
49000 infants followed for 7 years.
z
z
Apgar scores recorded at 1 and 5 minutes (and
for infants with AS<8 at 10, 15 & 20 minutes)
Histories recorded at 4, 12, 18 months and then
yearly to age 7
Dead
'Exposed'
Total
42
Not
dead
80
43
302
345
85
382
467
122
(Apgar score 0 - 3)
'Not exposed'
(Apgar score 4 - 6)
Total
Relationship between 10-minute Apgar scores and risk of death in the 1st
year of life among children with birth weight of at least 2500g
Risk in ‘exposed’
=
Risk in ‘less’ exposed
=
Relative risk
=
Risk difference
=
Attributable risk % = RD/ R(exp) =
?
?
?
?
?
Dead
'Exposed'
Total
42
Not
dead
80
43
302
345
85
382
467
122
(Apgar score 0 - 3)
'Not exposed'
(Apgar score 4 - 6)
Total
Relationship between 10-minute Apgar scores and risk of death in the 1st
year of life among children with birth weight of at least 2500g
Risk in exposed
= 42 / 122 = 0.344 = 34.4%
Risk in less exposed = 43 / 345 = 0.125 = 12.5%
Relative risk
= 0.344 / 0.125 = 2.8
Risk difference
= 34.4 - 12.5 = 21.9%
Attributable risk % = RD / R(exp) = 0.219 / 0.344 = 64%
Bad outcome eg death, h'capped infant
If exposure good/protective
If exposure harmful
(folate supplements, high AS
(foetal distress, smoking)
<1
1
>1
Relative risk
If exposure
harmful
If exposure good
Good outcome eg cure,
'normal' infant
z
z
55% of cerebral palsy (CP children had
1 minute scores >= 7
99 survivors with scores 0 - 3 (at 10, 15 or
20 minutes, all birth weights)
z
z
80% free of major handicap
12 had CP
11 mentally retarded, 1 normal
z 10 had had seizures on day of birth
z
z
1 in 8 asphyxiated neonates have CP and
developmental delays
z
<< 1in 8 of those without seizures
What is key difference between RCT and
cohort study?
z
RCT
z
z
Random allocation of intervention/exposure by
investigator
Cohort
z
Self selection of exposure/risk factor