IN PUBLIC HEALTH
VITRI WIDYANINGSIH
Attributable Risk (AR)
AR = (Risiko pada kelompok terpapar) - (risiko pada kelompok tidak terpapar)
◦ Berapa banyak risiko pada kelompok terpapar disebabkan oleh paparan, (bukan oleh risiko baseline)
◦ Risiko baseline secara teoritis dapat kita ketahui dengan mencari risiko pada kelompok tidak terpapar
Contoh:
◦ Risiko pada kelompok terpapar = 3/100
◦ Risiko pada kelompok tidak terpapar = 1/100
◦ AR = (3/100) – (1/100) = 2/100
◦ Walaupun risiko pada kelompok terpapar sebesar 3/100, 1/100 merupakan risiko baseline,
sehingga risiko yang disebabkan (attributable) oleh paparan adalah 2/100
Attributable Risk
Excess Risk =
Risk Attributable Risk
Attributable Risk =
(Risiko pada paparan+) –
(Risiko pada paparan -)
Relative Risk =
(Risiko pada paparan +) /
(Risiko pada paparan -)
risiko baseline
paparan
Attributable Risk
Excess Risk =
Attributable Risk
Risk
Attributable Risk = 0.78 – 0.5 = 0.33
risiko baseline
paparan
Relative vs. Attributable Risk
Relative risk merupakan ukuran yang penting untuk menandakan
kekuatan hubungan
◦ Penting untuk menentukan hubungan etiologis/ causal
2% attributable risk, 3% total risk artinya 66% total risiko pada kelompok terpapar
disebabkan oleh paparan
Attributable Risk Fraction
Excess Risk =
Attributable Risk
Risk
Attributable Risk =
(risiko pada paparan +) –
(risiko pada paparan -)
risiko baseline
paparan
Attributable Risk Fraction
Ukuran absolut
Excess Risk =
Attributable Risk
Attributable Risk = 0.78 – 0.5 = 0.33
Risk
33% penyakit pada kelompok
terpapar (baik yang sakit atau tidak)
disebabkan oleh paparan
Relatif terhadap risiko
pada kelompok terpapar
group
paparan
risiko baseline
Contoh (berkelompok)
Disease + Disease -
Exposure + 6 94 100
Exposure - 2 98 100
8 192 200
Population Attributable Risk
Seberapa besar penyakit yang ada di populasi disebabkan oleh paparan?
Mempertimbangkan prevalensi paparan pada populasi
Excess Risk =
Attributable Risk
Risk
Attributable Risk =
(risiko pada paparan +) –
(risiko pada paparan -)
risiko baseline
paparan
Population Attributable Risk (PAR)
Contoh
◦ risiko pada kelompok terpapar=3/100, risiko pada kelompok tidak terpapar=1/100
◦ prevalensi paparan=20%
◦ PAR = AR x prevalensi paparan
◦ PAR = (2/100) x 20%
◦ PAR = 0.4%
◦ “ Risiko kanker paru yang disebabkan oleh merokok pada populasi adalah 0.4% (baik terpapar maupun
tidak), bila 20% populasi terpapar rokok.”
Population Attributable Risk Fraction
Berapa % penyakit pada populasi disebabkan paparan?
Ukuran absolut di
Excess Risk = tingkat populasi
Attributable Risk
Risk
Population Attributable Risk =
PAR = AR x prevalensi paparan
% risiko di populasi
risiko baseline
paparan
Population Attributable Risk Fraction
(PARF)
PARF = PAR / total risiko
◦ risiko pada kelompok terpapar=3/100, risiko pada kelompok tidak terpapar=1/100
◦ prevalensi paparan=20%
◦ PAR=(2/100) x 20% = 0.4%
◦ Total risiko = total risiko pada populasi (terpapar+ tidak terpapar) = (3 + 1) / (100 + 100) = 4/200 =
2%
◦ PARF = PAR / Total risiko
◦ PARF = 0.4% / 2% = 20%
◦ 20% dari seluruh penyakit yang ada di populasi disebabkan oleh paparan
or
◦ If we could remove this exposure from this population, we would remove ~20% of the risk of this
disease
PAR and PARF
prevalensi paparan = 30%
Absolute measures
Excess Risk =
Attributable Risk PAR = 0.33 * 0.5 = 0.066
Risk
6.6% of disease in the population is
attributable to the exposure
Relative to risiko
pada kelompok
terpapar group
PARF = 6.6%/64%
PAR = RD x Prevalence
RD = 0.21 – 0.07 = 0.14
PAR = 0.14 x 0.5 = 0.07
PARF = 0.07 / 0.14 = 50%
Group Practice
Null value=0
More practice compute RD and RR
Flu Flu Total
Getting Yes No
Vaccine
Yes 5 95 100
No 40 60 100
Risk of getting flu for people who were vaccinated : 5/100 = 0.05
Risk of getting flu for people who were not vaccinated : 40/100 = 0.40
“The risk of Flu in those vaccinated
RD = 0.40 – 0.05 = 0.35 decreased by 35% as the risk of Flu in
people who were not vaccinated”
Ref: http://www.cche.net/usersguides/ebm_tips.asp
Relative Risk Reduction (RRR)
RRR = 1 – RR
How much less likely will someone get the disease if they are treated
✔ i.e. used more for beneficial exposures like treatments
Example:
◦ Risk in treatment group = 20/100 = 0.02
◦ Risk in control group= 80/100 = 0.08
◦ RR= (20/100)/(80/100)=0.25
More practice compute RRR
Flu Flu Total
Getting Yes No
Vaccine
Yes 5 95 100
No 40 60 100
Risk of getting flu for people who were vaccinated : 5/100 = 0.05
Risk of getting flu for people who were not vaccinated : 40/100 = 0.40
“The risk of Flu in those vaccinated is
RR = 0.05/0.40 = 0.125 0.125 times as the risk of Flu in people
who were not vaccinated group”
Getting Yes No
Vaccine
Yes 5 95 100
No 40 60 100
Risk of getting flu for people who were vaccinated : 5/100 = 0.05
Risk of getting flu for people who were not vaccinated : 40/100 = 0.40
“The risk of Flu in those vaccinated
RD = 0.40 – 0.05 = 0.35
decreased by 35% as the risk of Flu in
people who were not vaccinated”
(86/104320 = 0.00082)
New example:
◦ Risk in treatment group = 10/100
◦ Risk in control group = 20/100
◦ ARR=RD = (20/100) - (10/100) = 10/100
◦ Since each person taking this treatment reduces their risk by 10%, if 10 people are treated, on average, we will
prevent 1 case of the disease
◦ So, NNT=10
Number Needed to Harm (NNH)
Same thing as NNT— for harmful effect
◦ Commonly used in adverse medication events
New example:
◦ risiko pada kelompok terpapar= 45/100
◦ risiko pada kelompok tidak terpapar = 20/100
◦ RD=ARR= (45/100) – (20/100) = 25/100 = 0.25
◦ NNH = 1/ARR = 1/0.25 = 4
◦ Since each person who was exposed increases their risk by 25%, if 4 people are exosed , on average, we
will cause 1 case of the disease
◦ NNH = 4