BlankTemplate - University of South Alabama



BIOSTATISTICS

Measures of Diagnostic Test Performance

|  |Disease Present |Disease Absent |

|Test positive |A |B |

|Test negative |C |D |

Sensitivity = A/(A+C)

Specificity = D/(B+D)

Positive Predictive Value = A/(A+B) Dependent on population Dz prevalence

Negative Predictive Value = D/(C+D) Dependent on population Dz prevalence

Positive Likelihood Ratio = A(B+D)/B(A+C) Independent of population Dz prevalence

Negative Likelihood Ratio = C(B+D)/D(A+C) Independent of population Dz prevalence

Accuracy = (A+D)/(A+B+C+D)

Relative Risk (Risk Ratio): The incidence in exposed individuals divided by the incidence in unexposed individuals.

Relative risk is usually determined in a cohort study: A group of patients who have variable exposure to a risk factor are followed over time for an outcome. The cohort is identified before the appearance of the outcome under investigation; i.e., the cohort cannot be defined as a group of people who already have the disease.

Odds Ratio (Hazard Ratio): The odds that an individual with a specific condition has been exposed to a risk factor divided by the odds that a control has been exposed.

Odds ratio is used in case-control studies: Patients with a disease are identified and compared with matched controls for exposure to a risk factor.

Confidence Interval (CI): The range about a measured value that is estimated to contain the “true” value, to a specified degree of certainty. The boundaries of a confidence interval are often specified as 95% (2 standard deviations). The calculation of a confidence interval considers the standard deviation of the data and the number of observations. CI narrows as the number of observations increases. CI narrows as the variance (dispersion) of the observation decreases. A narrow confidence interval is good.

Margin of Error: Half the confidence interval.

Type I Error (Alpha Error): The probability of incorrectly concluding that there is a statistically significant difference in a dataset. Alpha is the number after a p-value.

P-Value: The probability that the observed difference in a dataset could have occurred by chance. P-value is a measure of the effect of chance within a study, not the probability that the result of the study is true.

Confidence Level: The probability that the observed difference in a dataset did not occur by chance; 1 – alpha.

Type I (Alpha): Rejecting the null hypothesis when the null hypothesis is TRUE.

In a study comparing a new drug to placebo, a Type I error would allow an ineffective drug onto the market.

Type II Error (Beta Error): The probability of incorrectly concluding that there is not a statistically significant difference in a dataset. This error often reflects insufficient power of the study.

Power: The ability of a study to detect a true difference; 1 – beta. Negative findings in a study may mean the treatment really did no good, or that the study was underpowered to detect a difference, most commonly due to not having enough subjects in the study.

Type II (Beta): Accepting the null hypothesis when the null hypothesis is FALSE.

In a study comparing a new drug to placebo, a Type II error would keep an effective drug off the market.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download