EpiMax Table Calculator

Epidemiology & Lab Statistics from Study Counts
With Chi Square, NNT & "Cost to Treat" Estimates

[For Demonstration Only-Not for Official Use]
Clinical & Economic Software Solutions    
Health Decision Strategies, LLC Princeton, New Jersey  USA


Data Entry Page

Instructions: Using 2x2 study data, you can change the "Title" and  fill in the four center cells in the table below (the cells in blue)  and if you wish, enter a "Cost Per Person" value. Hit the "Calculate" button to see the estimated results.

(Results generated will appear in the boxes outside and below the center cells.)
   

Target Disorder or Outcome

  
  Analysis Title:

Present

Absent

  
 

Case

Control

  
   

True Posititve(a)

False Positive(b)

a + b

Rx
Dx
H		 Control Group
Diag. Test positive
Exposed to Risk Factor

         
   

False Negative(c)

True Negative(d)

c + d

Rx
Dx
H		 -- Experimental Group
-- Diag. Test negative
-- Not Exposed to Risk

         
  Incremental 
Cost Per Person (CPP) 
Per Duration

a + c

b + d

a+b+c+d

  $

After you enter values into the four blue center cells above,
press the "Calculate" button, to see the estimates. 
 

 

     

Other names:

Sensitivity

a/(a+c)

True positive rate

Specifity

d/(b+d)

True negative rate

Likelihood Ratio +

 a/(a+c)/(b/b+d)

sensitivity/(1-spec)

Likelihood Ratio -

 c/(a+c)/(d/b+d)

(1-sensit)/specificity

  False positive rate

 b/(b+d)

1-specificity

False negative rate

 c/(a+c)

1-sensitivity

Probability of disease

 (a+c)/(a+b+c+d)

Prevalence

Predictive value positive

 a/(a+b)

Bayes' Theorem *

p(pos test wrong)

 b/(a+b)

 

Predictive value negative

 d/(c+d)

 

p(neg test wrong)

 c/(c+d)

 

p(test positive)

 (a+b)/(a+b+c+d)

Will test be positive?

p(test negative)

 (c+d)/(a+b+c+d)

Complement of P(tp)

Overall accuracy

 (a+d)/(a+b+c+d)

Probability correct

Relative Risk

 (a/(a+b))/(c/(c+d))

Risk without Rx / Risk with Rx

Odds Ratio

(a/c) /(b/d) = (ad)/(cb)

See note below

Attributable Risk

 (a/(a+b))-(c/(c+d))

Risk with exposure minus risk without.

NNT

 1/(a/(a+b))-(c/(c+d))

Number needed to treat

Cost to Treat

CPP x NNT

Cost for one improvement 
 
Chi Square
d.f.
p
Statistic
Value
Low 95% CI
Hi 95% CI
Odds Ratio
Relative Risk
Kappa
Sensitivity
Specificity
Pos Pred Val
Neg Pred Val

Chi-Square Statistics and Confidence Intervals

The Chi-Square section above estimates various additional statistics from a 2-by-2 table. It estimates a Yates-corrected chi-square, along with confidence intervals for other quantities relevant to two special kinds of 2-by-2 tables:
  1. analysis of risk factors for unfavorable outcomes (odds ratio, relative risk)
  2. analysis of the effectiveness of a diagnostic criterion for some conditions (sensitivity, specificity, positive predictive value, negative predictive value)

These concepts are explained in detail in an online Evidence-based Medicine Glossary. Confidence intervals for the estimated parameters are computed by a general method given in: Statistical Methods for Rates and Proportions (2nd Ed.) by Joseph L. Fleiss (Pub: John Wiley & Sons, New York, 1981).

Relative Risk: If 2x2 data is from a cohort study (prospective), then use the Relative Risk value which is the ratio of the rate of events in the control (unexposed) group "a/a+b" to the rate of  events in the experimental (exposed)  group "c/c+d".

Odds Ratio: If 2x2 data is from a case-control study, then use the Odds Ratio value which is the ratio of the odds of  disease for the experimental group over the odds of disease for the control group.  Or the ratio of the odds of being exposed or unexposed in the "case" group "a/c" to the odds of being exposed or unexposed the "control" group "b/d". 
Health Decision Strategies, 1999 - 2004
Model developed from the R. Hamm Calculators and Roberto Bissoni's Xpert  program.

[For Demonstration Only - Beta Testing - Not for Official Use]

HDS Home Page