Perfect & Sample Information
What would you pay to know if your decision is the right one? That's the real question you're asking yourself when you're asking how much information is worth. There are two ways of looking at it.
The expected value of perfect information, or EVPI, is a theoretical number that says how much a business should pay to know with certainty the outcome of a decision. Perfect information involves knowing for sure what will happen, and that's always theoretical; there's no way to know for sure what will happen.
Take the new blood test. The EVPI would be the price you would pay to run a study that would tell you exactly how accurate and how expensive the new test will be for your hospital. But since there's always some variance, EVPI is really just hypothetical.
On the other hand, the expected value of sample information, or EVSI, is a real number that says how much actual information is worth. That is, it will tell you how much money you should pay for results from an actual test, survey, or other form of data collection.
The EVSI, then, tells you how much you should pay to do the test run in your hospital. Trying the new blood test out won't provide perfect information because when you move from the 100 patients in the test to the whole hospital, it could result in other problems, but it will give you more information than you have now.
So, how do you calculate EVSI? To start with, you need to know how much accurate and inaccurate information is worth. There are several ways to calculate this, including with decision trees and payoff tables, but essentially, what you want is to know how much benefit there is in using the information from the test run. If that information is accurate and things turn out the same way in the whole hospital that they did in the one department, that benefit will be one value. If the information is inaccurate (if, say, the test run shows that the new blood test is inexpensive, but it becomes expensive when rolled out to the whole hospital), then the number will be different.
The formula for EVSI is:
EVSI = (Accurate information + Cost of information) - Inaccurate information
For example, let's say that you figure out that having accurate information from the test run will be worth $50,000 to the hospital. Inaccurate information from the test run will only be worth $20,000. Meanwhile, the actual cost of the test run is, say, $15,000.
You can calculate your EVSI:
($50,000 + $15,000) - $20,000 = $45,000
Thus, $45,000 is the expected value of your sample information, or how much benefit your hospital will derive from doing the test run.
Information can help businesses make better decisions, but it comes at a cost. The expected value of perfect information, or EVPI, is a theoretical number that says how much a business should pay to know with certainty the outcome of a decision. On the other hand, the expected value of sample information, or EVSI, is a real number that says how much actual information is worth. Both are ways of estimating how much a company should pay for additional information while trying to make a decision. The formula for EVSI is EVSI = (Accurate information + Cost of information) - Inaccurate information.