Messages, Structure and Prior Authorization
National Committee on Vital and Health Statistics
Subcommittee on Standards and Security
August 17, 2004
MediMedia Information Technologies
On behalf of MediMedia USA, it is my pleasure to testify on formulary standards to the National Committee on Vital and Health Statistics (NCVHS) subcommittee on standards and security, and make myself available for questions and answers.
As I testified in July, we distribute the InfoScan Formulary Database™, which contains more than 3,400 health plan, PBM, PPO and self-insured employer formularies. In addition to most of the plans associated with Rx Hub and CAQH, we represent many of the smaller plans and PBMs who have thus far chosen not to affiliate with those organizations.
As I also testified, we have been providing a formulary database to electronic health records (EHR) and ePrescribing software companies since 1997. Our clients include some of the ePrescribing and EHRs that have testified to this subcommittee, in addition to WebMD’s Medical Manager, GE Medical’s MedicaLogic, Cerner, NextGen, Misys and others – a veritable a “who’s who” of mature health care information technology providers.
My goal is to represent the interests of all of these stakeholders in providing you with an alternate but very relevant viewpoint in the areas of formulary, structure and prior authorization.
As we testified in July, our health plan and PBM clients provide information to us in a variety of formats, including flat files, Excel, MS Word, email, phone calls and faxes. In the absence of a standard for collecting this data, we have become experts at doing so, validating the information and aggregating it into one standardized data set. A submission standard would be quite desirable, but not necessary for us.
Frequently the formulary decisions we receive are before they have been entered into the processors computer and what we communicate in our formulary information more closely resembles the intent of the P&T committee. That said, we would caution that, if not constructed intelligently, such a standard could limit plan creativity, putting a ceiling on the types and quantity of information. Therefore, we believe that submission formats should allow maximum flexibility.
In May, CAQH testified that they had developed a “standard” submission format. We’d be open to the committee recommending CAQH’s submission formats as a starting point; however, it is possible that they would have to be expanded, as some of our clients submit a great deal more information than I understand is submitted to CAQH. Such an expansion would allow for the previously mentioned need for flexibility.
Furthermore, I’d argue that a standard submission format should be developed and maintained by an ANSI-accredited standard development organization (SDO), so that all stakeholders would have the opportunity to provide input. NCPDP is a logical SDO with input from the physician community.
Outgoing Data Sets
We collect, validate, aggregate and maintain this rich set of data, providing it to clients on a weekly or monthly schedule, depending on what best meets the software company client’s individual needs. We distribute our database via CDRom or FTP, again at the client’s request. Real time interactive transmission is an area we’re piloting to support some specific installations.
We maintain a standard format for ease of comprehension and implementation; however, in some cases, special requirements by customers have necessitated the need to customize the data.
As I testified in July, we think that the outgoing data set could be standardized to minimize custom coding requirements. We would be happy to provide our formats as a starting point, or work with any of the other formulary aggregators’ formats.
Factors that you might consider when thinking about this recommendation:
(1) Because of reasons I stated earlier, we understand that our data is much broader than any other aggregators’ data so, if you used our formats, we’d cast a wider net, if you will.
(2) We did not use an existing standard to create our database, mostly because there was no standard at the time. Nevertheless, we do not believe it would be difficult to modify our data sets to be NCPDP or HL7 compliant.
(3) We also recommend that any standards minimize disruption to the vendors who have taken the lead in this market and not penalize them for the leadership position they have taken.
Formulary Data & Messages
In addition to drug status and tiers, we have step therapy rules, relative price ranking for on-formulary drugs, policies, rules or criteria for prior authorization, therapeutic guidelines and advisory and restrictive notes.
The content and breadth of a formulary as well as definition of terms vary by plan. This creates a challenge for EHRs and ePrescribing companies, just as it has for the processors who have had to try to adopt those new definitions into their legacy claims systems. While there is a benefit in terms standardization we don’t want it to dampen the creativity and competition between plans.
Nevertheless, this will be a tricky process that will require input from all stakeholders. To keep costs in check while improving outcomes, it is important that the plans and PBMs have maximum flexibility to creatively customize a formulary for their respective situation or client base.
Developing formulary data and messaging standards should be lead by the plans, with input from other stakeholders. While it isn’t an ANSI-accredited SDO, an organization such as AMCP could be an entity that could be part of the process. Perhaps they could collaborate with a transaction SDO such as NCPDP.
We agree with the testimony at the July hearing that one of the biggest challenges in ePrescribing is obtaining the patient’s demographic information from the practice management system. However, getting the information into the EHR or ePrescribing application only overcomes one of the barriers.
The other is linking the formulary with the health plan card information obtained by the practice. You see, often an employer will self-insure for the prescription benefit, contracting directly with the PBM. In this case, the patient will have separate health and pharmacy cards. The office staff will collect only the information required to submit the medical claim – the information on the health plan card – which may not be the information necessary to link the patient to the formulary. (Of the tens of million United HealthCare Group patients, 25% have separate PBM coverage.)
Rx Hub has worked around this by creating a hub and master patient index; however, not every PBM or plan feels it is in its best interest to join Rx Hub and I am questioning forcing such affiliations. If there were a standard way of identifying health plans and displaying this number and a formulary number on the plan cards, this would facilitate linking a specific formulary to a plan. While there would clearly be challenges, stakeholders working through a SDO could work out the details on how this would be accomplished. This would also aid patients in understanding the differences between formularies when selecting a benefit.
Pre or Prior authorization is, of course, the process of obtaining certification or authorization from a health plan for specified medications or specified quantities of medications. It often involves appropriateness review against pre-established criteria. Those criteria can vary by plan and, within a plan, by drug. The number of medicines with that require prior authorization vary greatly between plans.
The process of obtaining approval is onerous, by design. It’s purpose is to encourage appropriate use of medications most likely to have certain risk factors, and the approval criteria is generally developed and endorsed by the plan’s P&T committee, based on information from the FDA and manufacturers, medical literature, actively practicing consultant physicians and appropriate external organizations.
Failure to obtain prior authorization often results in a financial penalty to the patient, so physicians are highly reluctant to prescribe those drugs thus labeled. In fact, almost any physicians’ office that has even a moderate number of managed care patients will tell you that prior authorization tops its list on a “pain scale.”
For this reason, the ePrescribing system that can reduce the “pain” of prior authorization will be making a substantial positive impact on a practice.
I also believe that as ePrescribing becomes more commonplace, the rate of on-formulary prescribing will increase, making prior authorization a more attractive cost-containment tactic. Automating the process will allow clinically appropriate prescribing.
In today’s paper world, the prescriber does not know if the drug is on prior authorization or not. While he or she quickly learns that it’s likely that growth hormones or anti-fungal agents have been designated as requiring prior auth, what trips him or her up are therapeutic categories that are less consistent across plans. Examples are Cox-2s such as Celebrex, Vioxx and Bextra, or Proton Pump Inhibitors where availability of lower cost options has created prior authorization restrictions on many medications.
Should the office want to continue with a prior authorization request, the staff would obtain a form from the plan or a Web site. The form has a series of questions designed to help a clinician determine if the prescription is medically necessary. While it is more complex than “yes/no”, according to our analysis, the example that Medco presented in July is a most complex example.
The fact is, computers were designed to automate paper processes like this. At a minimum, when the prescriber is using a software solution that leverages a Formulary Database, these drugs will be flagged as requiring prior auth. The physician can access notes and restrictions on that drug and make the decision as to whether to put the practice through the “pain” of requesting a formulary exception.
But that’s only a first step.
An algorithm can run either in the software system or interactively that allows the physician to enter diagnosis codes, answer questions and document his/her clinical judgment. Some plans for some drugs might issue an approval code at this moment. In other cases, a form would be created and transmitted to the plan’s clinicians for approval. When approval is obtained, the code can be transmitted with the prescription to the pharmacy, where it can be included with the claim transmitted to the prescription payor.
We volunteer to do a demonstration project with one of our software and health plan clients whereby we would create an algorithm for the prescriber, either giving him or her a response within seconds or transmitting a completed request to the plan for approval.
- The biggest challenge with formularies is linking the correct one to the patient. A standard formulary identifier included in the plan ID would resolve this issue.
- Formulary information at the plan level provides enough granularity to satisfy patients and prescribers needs. Further benefit information is not precise, nor provides significant extra value
- Computers were designed to automate processes like prior authorization.
Why do you only distribute your formularies on a weekly or monthly schedule?
Because we have so many formulary suppliers, it would be cost-prohibitive and logistically impossible to distribute our database after each formulary update, which is, as you know, generally after the plan’s P&T committee meets.
You said that the content and breadth of a formulary varies by plan or PBM. Can you provide some examples?
Sure. Definitions of basic terms such as “preferred status” are different, and plans do not manage tiers uniformly. Sometimes in a two-tier formulary, we find brands in tier one. Other times, we find expensive generics in tier two. Managing this is a challenge for EHRs and ePrescribing companies.
You said that growth hormones commonly require prior authorization but the process is unusually onerous. What are the most common therapeutic categories that require PA?
While growth hormones are one of the four most common therapeutic categories with prior authorization, they were only 5% of all drugs thus designated. The other three most common therapeutic classifications for prior authorization were Cox-2 inhibitors, impotence agents and anti-fungals.
In July, there was testimony that only 2% of drugs were under prior authorization. Is that consistent with your database?
We analyzed our database and found that nearly half of our HMO plans had fewer than 50 drugs labeled as prior auth. Thirteen percent had more than 100 products under prior auth, and 4% had more than 200. It’s the more aggressive plans that have more than 100 products under prior auth that presents the greatest challenge for the prescriber.