When creating a data mining run, you can specify that Proportional Reporting Ratio (PRR) and Chi-square should be computed and included in the results for each drug-event combination against a specified background database.
Before you use the PRR calculator, you must specify three case series (based on the same configuration) to represent a background database, the events of interest, and the drugs of interest. For example, to compute PRR for combinations of analgesics and respiratory ailments in women over age 65, use a case series that includes cases for women over age 65, a case series that includes cases for analgesics, and a case series that includes cases involving respiratory ailments.
1. Click the Case Series tab.
The Case Series page appears.
2. Identify or define the three case series to use.
3. Click PRR Calculator.
The PRR Calculator page appears.
4. From the Configuration drop-down list, select a data configuration.
5. For the following types of case series, click Browse and select a case series. You must base all three case series on the same configuration.
6. To define the chi-square calculation method, check or clear Apply the Yates correction for chi-square.
7. To define the PRR calculation method, check or clear Include drug of interest in the comparator set.
8. Click Calculate.
The observed counts and results appear.
Note: When PRR is computed by a data mining run, a, b, c, and d represent counts of drug-event combinations by default. You can select an option to count cases with those combinations instead. However, the PRR calculator always counts cases.
|
Cases |
Specific drug |
All other drugs |
|
Specific event |
a |
b |
|
All other events |
c |
d |
PRR = (a / (a + c)) / (b / (b + d))
If b=0 (and Include drug of interest in the comparator set described below is cleared), then the formula is adjusted as follows to avoid the possibility of division by zero:
PRR = ((a+0.5) / ((a+0.5) + (c+0.5))) / ((b+0.5)/((b+0.5) + (d+0.5)))
If you checked Include drug of interest in the comparator set, the application computes PRR as:
PRR=(a / (a + c)) / ((a + b) / (a + b + c+ d))
|
Cases |
Specific drug |
All other drugs |
|
Specific event |
E(a) = ((a+b)(a+c))/(a+b+c+d) |
E(b) = ((a+b)(b+d))/(a+b+c+d) |
|
All other events |
E(c) = ((c+d)(a+c))/(a+b+c+d) |
E(d) = ((c+d)(b+d))/(a+b+c+d) |
The chi-square of PRR (PRR_CHISQ) for the combination of a particular drug and particular event is computed as follows, where E represents the expected count under the assumption of no relationship between drugs and events:
PRR_CHISQ = ((a-E(a))2)/E(a)+ ((b-E(b))2)/E(b)+ ((c-E(c))2)/E(c)+((d-E(d))2)/E(d)
If you checked Apply the Yates correction for chi-square, PRR_CHISQUARE is computed as follows:
PRR_CHISQ = ([max(0, |a - E(a)| - 0.5)]2)/E(a) + ([max(0, |b - E(b)| - 0.5)]2)/E(b) + ([max(0, |c - E(c)| - 0.5)]2)/E(c) + ([max(0, |d - E(d)| - 0.5)]2)/E(d)
Note:
|a - E(a)| = |b - E(b)| = |c - E(c)| = |d - E(d)| for every 2x2 table; this formula can also be expressed as:
PRR_CHISQ = (1/E(a) + 1/E(b) + 1/E(c) + 1/E(d)) [max(0, |a - E(a)| - 0.5)]2
The P-values for chi-square values are also computed. The P-value is the probability that chi-square would be as large as the value for PRR_CHISQ by chance alone if there were no causal relationship or consistent association between the drug and the event.
Note: If any of E(a), E(b), E(c), or E(d) is zero, the chi-square and P-value cannot be computed and are defined arbitrarily as -1.
9. In the table for observed counts, you can click the value of N in a column to display a menu. Click View Cases to drill down to cases that have the specified drugs and specified events.