How to calculate Odds ratio, relative risk and 95% confidence interval by two-by-two table?

You can determine statistically significance with Odds ratio, relative risk and 95% confidence interval by two-by-two table which treats such binomial values which called as outcome as true or false, for example, enabled or disabled and death or survival and such binomial values which called as predictor as positive or negative, for example, with intervention or without, with exposure risk factor or without and active drug or placebo. 

Odds ratio (OR) is used for case control study. When OR is 1, the predictor has no effect. When OR is greater than 1, the predictor has positive effect to outcome. If it is smaller than 1, the predictor has opposite effect to outcome.

Relative risk (RR) is used for prospective cohort study and randomized controlled trials. It means that when RR is greater than 1 the predictor has positive effect to outcome and when RR is smaller than 1 the predictor has negative effect to outcome.

If 1 is between lower bound and upper bound of 95% confidence interval of Odds ratio and relative risk, it is not statistically significant.

  TRUE FALSE Marginal total
POSITIVE a b a + b
NEGATIVE c d c + d
Marginal total a + c b + d N 

You could calculate Odds ratio and 95% confidence interval with formula below;

\displaystyle OR=\frac{\frac{a/(a+c)}{c/(a+c)}}{\frac{b/(b+d)}{d/(b+d)}}=\frac{ad}{bc}

 

\displaystyle 95\%C.I.=\exp\left(\ln OR \pm 1.96\sqrt{\frac{1}{a}+\frac{1}{b}+\frac{1}{c}+\frac{1}{d}}\right)

You could calculate relative risk and 95% confidence interval with formula below;

\displaystyle RR=\frac{a/(a+b)}{c/(c+d)}

 

\displaystyle 95\%C.I. = \exp\left(\ln{RR}\pm1.96\sqrt{\frac{1}{a}+\frac{1}{c}-\frac{1}{a+b}-\frac{1}{c+d}} \right)

四分表からオッズ比と相対危険度および95%信頼区間を計算する

 有効か無効か,生存か死亡かなどの値が TRUE か FALSE の二項分布する名義変数をアウトカムとし,危険因子の暴露ありかなしか,介入ありかなしか,実薬か偽薬かなどの値が POSITIVE か NEGATIVE の二項分布する名義変数を予知因子とする場合,四分表により予知因子がアウトカムに及ぼす影響が統計的有意か否かをオッズ比や相対危険度の95%信頼区間により判定できます.

 オッズ比は症例対照研究に用いられ,オッズ比が1だと予知因子は何の効果もないことになります.オッズ比が1より大きいと予知因子によるアウトカムの発生が増えることになります.逆に1より小さければアウトカムの発生を抑制することになります.

 相対危険度は前向きコホート試験やランダム化比較試験で用いられ,1より大きければ予知因子はアウトカムの発生を高め,1より小さければ予知因子はアウトカムの発生が低くなることになります.

 オッズ比,相対危険度共に95%信頼区間が1を跨いでいる場合には統計的有意とはいえません.

  TRUE FALSE Marginal total
POSITIVE a b a + b
NEGATIVE c d c + d
Marginal total a + c b + d N 

オッズ比ORは以下の式で求まります.

\displaystyle OR=\frac{\frac{a/(a+c)}{c/(a+c)}}{\frac{b/(b+d)}{d/(b+d)}}=\frac{ad}{bc}

オッズ比の95%信頼区間は以下の式で求まります.

\displaystyle 95\%C.I.=\exp\left(\ln OR \pm 1.96\sqrt{\frac{1}{a}+\frac{1}{b}+\frac{1}{c}+\frac{1}{d}}\right)

相対危険度RRは以下の式で求まります.

\displaystyle RR=\frac{a/(a+b)}{c/(c+d)}

相対危険度の95%信頼区間は以下の式で求まります.

\displaystyle 95\%C.I. = \exp\left(\ln{RR}\pm1.96\sqrt{\frac{1}{a}+\frac{1}{c}-\frac{1}{a+b}-\frac{1}{c+d}} \right)