Calculating Confidence Intervals



At the beginning

First enter the pretest data, and calculate the standard deviation and initial 95%-confidence interval for the pre-test probability. If you don't want to calculate confidence interval, leave the number of cases box empty.
Enter the following data:
Pretest Probability: Number of Cases:
Standard Deviation for pretest probability:
Confidence interval for pretest probability: [ , ]
or in percentages: [ % , % ]


Test


( If you don't want to use test in these calculations, uncheck the box above and procede to the Benefits and Risks section. )

Now enter the data for the test:
  • Sensitivity (the conditional probability of positive test result, given presence of the disease),
  • Specificity (the conditional probability of negative test result, given absence of the disease), and
  • the number of cases on which the data is based for both of them.
The program will calculate Standard Deviation for Specificity and Sensitivity.
If you wish to use your own estimate of the error, check the box located next to the desired standard deviation entry and enter your estimate into the box. If you want to evaluate error instead, don't forget to uncheck the box and enter the number of cases.
Enter the following data:
Sensitivity:
No. of Patients:
Specificity:
No. of Patients:
Std. Dev. for Sensitivity: Std. Dev. for Specificity:
False Negative: False Positive:
Positive Likelihood Ratio LR+:
 Negative Likelihood Ratio LR-:
Std. Dev. for LR+: Std. Dev. for LR-:


Posttest Probabilities

Conditional probability of the disease, given that the test result is positive.
P ( D+ | T+ ) =
95% confidence interval : [ , ]
or in percentages: [ % , % ]
Conditional probability of the disease, given that the test result is negative.
P ( D+ | T- ) =
95% confidence interval : [ , ]
or in percentages: [ % , % ]


Benefits and Risks of the Treatment

To calculate the thresholds, we need benefits, risks and the standard deviations of the benefits and risks. You have two options:
  1. Either enter the Benefits (B), Risks (R) and their standard deviations associated with your treatment into the boxes from your sources;
  2. Or, click on "Calculate..." button to enter the utility functions, and to calculate Benefits and Risks, together with their standard deviations.
Benefits:
 St. Dev. for Benefits:
 Risks:
 St. Dev. For Risks:


Risk of the Test

Sometimes a test has an associated risk with it. In such a case, please enter the risk and it's standard deviation in the boxes below. If this test has no risk associated with it, leave the box empty, or enter a zero.
Risk of the test:
 St. Dev. For Risks of the test:


Thresholds

Finally, we will calculate the confidence intervals for the two thresholds:
  • The testing threshold: if the probability of the disease in our patient is below this number, do not administer test, and continue observing the patient.
  • The treatment threshold: if the probability of the disease in our patient is above this number, administer the treatment.
Benefit/Risk:
 St. Dev. for Benefit/Risk:
Testing Threshold:
= % 		Treatment Threshold:
= %
Confidence Interval:
[ , ]
or in percents: [ % , % ] 		Confidence Interval:
[ , ]
or in percents: [ % , % ]


"These are theoretical calculations. Any clinician seeking to apply or consult these calculations is expected to use independent medical judgement in the context of individual clinical circumstances to determine applicability of derived calculations".