Sentinel effect of prior authorization

[participant5]

Are there any estimates of the sentinel effect of prior authorization programs?
How can it be measured? How long might it last?

Finally, what is considered to be a reasonable threshold for PA approval rates considering the possibility of the sentinel effect? For instance, is it worthwhile to have a PA for a product when the approval rate is 95%?

[paul.polansky]

In response to your question, I do not know of any study that has been completed to measure the sentinel value of Prior Authorization (PA) but from experience can tell you that it exists.

The rate can vary greatly for a number of reasons including the type of plan, the market competitors, PhARMA strategies and provider relationship. The sentinel effect will usually remain as long as you do not stop managing a PA, once you no longer manage the PA the provider network will gradually increase the use of those medications you have removed from management, often the news being delivered immediately to providers by PhARMA representatives.

Whether you are looking at this in terms of keeping providers from arbitrarily submitting inappropriate information or to manage workflow once you put a process in place, educate providers, deliver consistent decisions and messages in a sensible quality driven manner, you will see that the majority of outlier requests will fall off.

Depending on the type of product and the reason for the Prior Authorization (formulary, quantity or dosing, clinical pathway for example) you will continue to see reductions in inappropriate requests over time, with a rate based on experience of about 20-30%.

The simple way to measure this is to look retrospectively at PA rates, aapproval and denial rates as a benchmark then going forward assign some reason to each request to measure on a periodic basis. Differentiate between those denials for clinical rules, administrative and other reasons to get a good picture.

PA rates for appropriate denial can be as high as 40% of requests but does vary dramatically by plan and medication, what you want to ensure is that the PA process is quality driven not just a roadblock to service. To that comment, you may want to continue managing a PA even when your approval rate is 95% because there may be significant patient risk if inappropriate or cost issues that have to be considered across the bigger picture.

[jeff.dunn]

I am also unaware of any data that reports the sentinel effect of prior authorization programs - this would be extremely difficult to measure because we don't have a way to measure physician behavior that is not captured in claim data. However, based on anecdotal interactions with providers, there is a significant impact. Most providers don't want the hassle if there are other clinically appropriate alternatives. Also, we have seen an increase in utilization after we have removed PAs in certain categories.

Reasonable thresholds vary depending on the disease state and the drug that carries the PA. A denial rate of 5% (or approval rate of 95%) may be appropriate if the drug is one that has a high likelihood of inappropriate use, is costly, etc. Our denial rates vary from 5-95%. We have removed some PAs (and step therapy edits, etc.) due to a low denial rate in categories that do not have good clinical alternatives (i.e., TNFs for RA), better contracting from pharma, etc.

Most plans do monitor the number of PA requests, and subsequently, the number of approvals and denials; and make appropriate changes to coverage.

[andrew.maiorini]

This is a very interesting topic to me as well. Often managers will look at the approval data and draw inappropriate conclusions. For example I often hear exactly what was mentioned at the start of this post "since we approve 95% we should remove this PA". As discussed the answer is much more complex than that. One needs to look at the entire picture before making any decisions based on approval %. If tha PA is appropriate and the requests only reach the PA department when they are also appropriate then there is no reason to remove the PA, it is working! What you would need to know is how often Physicians decided not to write for it unless it was clinically required and what increase can be expected once Pa is removed.

One additional factor is the rate of PA for a particular product in the market. For example you see this effect most often in markets where most managed care plans require PA for that product. Many physicians assume every plan requires PA if they encounter PA enough times. Sort of reminds me of the community protection provided by mass immunizations......