[This Transcript is Unedited]
Hubert H. Humphrey Building
Conference Room 705-A
200 Independence Avenue, S.W.
Washington , D.C.
PARTICIPANTS:
TABLE OF CONTENTS
Agenda Item: Call to Order, Welcome and Introductions
MR. REYNOLDS: Good morning. This is a meeting of
the Standards and Security Subcommittee of NCVHS. NCVHS is a statutory public advisory body on health information and policy for the secretary of health and human services. We are being recorded and are on the Internet. As we begin introductions, I will ask each of the members of the committee to note any possible conflicts of interest. We also will have a few of our members on the phone today, and I'll have them introduce themselves also.
I am Harry Reynolds, Blue Cross and Blue Shield of North Carolina. Co-chair of the committee with Jeff Blair and I have no conflicts.
Jeff.
MR. BLAIR: I'm Jeff Blair, Director of Health Informatics Loveless Clinic Foundation. I am co-chair of the Subcommittee on Standards and Security and I'm not aware of any conflicts with the subjects we're talking about.
Dr. Fitzmaurice: Michael Fitzmaurice, Agency for Health Care Research and Quality, liaison to the National Committee and staff to the Standards and Security Subcommittee.
DR. LEVINE: Randy Levine, Food and Drug Administration.
DR. NELSON: Stuart Nelson, National Library of Medicine.
MS. GOVAN-JENKINS: Wanda Govan-Jenkins, National Center for Health Statistics, CDC, Staff to the standards and security subcommittee.
DR. COHN: Simon Cohn, chair of the NCVHS, associate executive director for Health Information Policy for Kaiser Permanente, member of the subcommittee, and no conflicts of interest.
MS. TRUDEL: Karen Trudel, Centers for Medicare and Medicaid Services, sitting in for Denise Buenning as chief staff to the subcommittee.
MR. REYNOLDS: Justine?
DR. CARR: Justine Carr, Beth Israel Deaconess Medical Center member of the committee.
MR. REYNOLDS: Jorge?
DR. FERRER: Jorge Ferrer, Veterans Health Administration, staff to the subcommittee.
MR. REYNOLDS: Did any other members join on the phone?
DR. HUFF: This is Stan Huff. I just joined.
MR. REYNOLDS: If you'd introduce yourself please, Stan.
DR. HUFF: Stan Huff, with InterMountain Health Care at the University of Utah in Salt Lake, and no conflicts unless we talk about HL7 and I may have to recluse myself.
MR. REYNOLDS: Okay. We're going around the room now.
(Introductions around the room.)
Agenda Item: NLM Testimony— Mapping of RxNorm and NDC
MR. REYNOLDS: Okay, our first agenda item is an
NLM testimony mapping of RxNorm and NDC, and Stuart you have the floor.
DR. NELSON: You have to forgive me, I'm a Mac person, I gotta figure out how to run POWERPoint on this thing. How do I get to my slideshow? Ah, there it is.
MR. REYNOLDS: You should have been ready out of the box, Stuart.
DR. NELSON: Thank you. I wasn't quite sure what you wanted to hear about today, so I'm just going to try and give you a brief update of where we are with RxNorm with a lot of special attention to what we're doing with the NDC codes. My presentation is pretty short. So, if there are questions at any time please don't hesitate to interrupt. I'm sure we'll end up with lots of time and perhaps even an early break.
Well, what I was going to talk about was the experiences that I know about with Electronic Prescribing and RxNorm. Some problems that have come to my intention about NDC codes, then I'm going to talk about NDC codes in RxNorm. Our sources of them, how we represent them right now, what we're going to be planning to do in the future, and a few moments about outdated NDC codes. Just try to think ahead about where we're going to need to be going with our NDC code portion of our work.
Well, there have been the pilots that were sponsored, I believe, by CMS on Electronic Prescribing and at this point, we had not received any significant feedback from the pilots. There were a few small requests for assistance in understanding RxNorm, but I think RxNorm was fairly well documented, and people understand it pretty well. I will observe that as somebody who practiced medicine for a good number of years, that to me, the Electronic Prescribing and the issues around it are much more concerned with dispensing than they are with prescribing. That most of the dialogue relates to lots of things that have to do with dispensing medications, and very little has to do with the way a doctor actually writes a prescription. As a result of that, because of the focus on dispensing and the relations of dispensing to billing and so forth, a lot of the communications appears to be related or based on NDC codes.
Now there are problems with using NDC codes for prescriptions. First of all, there is no doctor in the country who knows the NDC codes that well. There are some drugs for which there might be as many as 250 NDC codes for a single drug. It is very hard to maintain knowledge about drugs at the NDC Code level. So if you want to do things like interaction checking or allergy checking, it is very hard to maintain that at the NDC Code level. You just have to go through and sort through it all. One of the things that you can't do because of the maintenance problem right now is to avoid the problem of overprescribing. Ya know, how do you know the patient is taking Endoral and Propranolol? You can't recognize that from an NDC Code. And this, I think, is a significant issue that needs to be addressed.
On the other hand, there are some real benefits in using NDC codes. Your cost, your reimbursement, your inventory data are all based on your NDC codes. And I think that there is a real value in doing that. There is a precision of expression that's allowed with NDC codes in that it allows you to specify a particular product including the excipients involved, whether or not something is-- say for example, alcohol free, or lactose free, so forth, that you can not necessarily represent well at the RxNorm level of expression. So there are some real benefits in expressing things at an NDC level.
There are some other problems involved. Recently, I've seen some work done on NDC codes that suggest that as many as 20% of NDC codes that are being handled through a centralized clearing house involve locally generated NDC codes. These could be repackaged drugs, drugs that have been compounded in the pharmacy there, or (and this one was quite a surprise to me!) there are NDC codes that are sometimes generated to cover the procedure involved in administering the drug, which seems to be fairly far afield from the idea of what you're supposed to be doing with an NDC Code. But in fact, that is what they are doing. There is some difficulty with representation of NDC codes, whether it be a 10-digit identifier with 3 fields separated by dashes, or whether it be 11-digit with no dashes or hyphens in it. You can go from a 10-digit to an 11-digit, but going from an 11-digit to the 10-digit, unless you know where that extra 0 was inserted, you're just guessing. So it's only a one-way conversion. You can't convert back. I would put forth the idea that the NDC code is not a meaningless identifier, although it is conjunction of three meaningless identifiers. The very fact that its 3 meaningless identifiers conjoined together has a certain semantics to it, and so it's not quite a meaningless identifier. The semantics that are presented are not helpful semantics, particularly. Currently, there is no central authority for assigning NDC codes. Now, I believe Randy is going to talk about a proposed rule where the FDA is going to establish a central authority for establishing NDC codes, and I think that that would be something to be commended and encouraged as much as possible.
Another thing I was asked to talk about was: How ready is RxNorm to be used for prescribing at a clinical level? Well, there are basically 3 different types of elements of drugs or drug products that could be prescribed at a clinical level. One of these are the drug delivery devices. These would be things like a kit containing more than one drug, a prefilled syringe like an EpiPen or of birth control pills where there's three different types of pills in the same package. We have around 1,800 of these in RxNorm right now. We do not have a standard name, and we do not have them linked into what the particular drugs are that are within them and what the particular ingredients of those drugs are and strengths. This is something we're currently working on in our maintenance and system. And we expect to have that implemented early in 2007, that we will have those where they will also be linked also to the brand and drugs and the generic drugs. Of the generic drugs, right now we have 17,600 generic drugs defined as ingredient, strength, and dose form as administered to the patient, and we have about 14,000 drugs with brand names. I think that these numbers have started to level off. In fact they actually dropped somewhat which was encouraging because that meant we had found some previously unrecognized synonymy where two things were being used as the name for the same thing.
One of the encouraging things is that we have
been able to be cut by working with the pharmacy knowledge based vendors and their vocabularies. We're doing our best to support interoperation between the users of the different vocabularies. One of the notable efforts right now is that of the Cheddar(?). Which where the Department of Defense which uses First Data Bank, I think because it's primary pharmacy knowledge base is communicating to the VA which has its own particular vocabulary. And we're working actively with them to facilitate that interoperation.
So where do we get our NDC codes that are in RxNorm right now? Well, as I pointed out before, there is no single authority. No single source is definitive. We're getting it from three different sources right now. We have the FDA NDC listing, that's the NDC database. We have the VA National Drug Formulary has NDC codes associated with it. And Multum of commercial vendors has been providing us with NDC codes. The other commercial sources have not chosen to share their NDC data with us. We would like to be able to at least see that, because I'm going to suggest that this is an important way of testing quality of looking to do quality assurance on that.
In the NDC listings, we have found various kinds of data anomalies that we've been pointing out and the current listings of the NDCs are of much higher quality then they were two years ago. We're all aware of the recent inspector general's report about obsolete drugs stolen in the NDC listings. And that is something of a concern, but I'm not quite sure how to go about approaching it at this time.
Our original plan with the NDC codes was to have them linked as attributes only to the generic drug. And we list them as attributes of these drugs at the level of the source so that if Multum has a given drug and says it has an NDC code, we will list those NDC codes associated with the drug as the source of it being multum. So, we have 336,000 occurrences of NDC codes. Now there are certainly a lot overlaps in that because the FDA listings and Multum listings and the VA listings all intersect quite a bit. What we have found is that we have about 20,000 conflicts. Out of that 20,000 conflicts; 12,000 were partly as a result of our changing philosophy about linking the NDC codes to the generic name and now some of them we are linking to the brand name. And as a result of that change in philosophy we have about 12,000 them, we have the same NDC code on a generic drug as on a brand name, which basically is the same ingredients and same strength. The other 8,000 provide us with a really good method of looking at quality control in RxNorm. Our editing, our model for how to represent the drug, to see whether or not we have places where we did not have clearly demarcated boundaries between how we're going to name things. Differences in dose forms and oral solution versus an oral suspension. One source may have listed it has a solution and another as a suspension with the same NDC code. So there are these kinds of issues that do arise. And some of them are just frightened disagreement between the sources. And since we don't have an authoritative source, there is very little we can do about that except to just say, “Well, we know there is a disagreement there.” and hope it can be resolved.
Well, as some of you may know, last November the
FDA and the National Library of Medicine entered into, well it had signed an agreement sometime before, but actually began the process of transmitting the labels for medications in electronic format. The FDA would send them to us in XML and that we put them on a web site and make them available for download at a site called the DailyMed. One of the things that's interesting about that is that the NDC codes are listed in the label under the house applied section in the label. And what we are looking is to use the DailyMed to replace the NDC listings with the FDA as our primary source of information from the FDA. And there are certain kinds of issues that arise for us in doing that. One of which is that you can have more than one label for the same product. If it's a different package or whatever, they may have different labels but the same product. So you can have the same NDC code in more than one label.
We have to learn how to recognize when a drug has
become obsolete or when a product has been withdrawn. If there is no label, we can say everything in that particular label has become obsolete if there is no label listing that drug. So, then we could then list this as being something that's obsolete. I haven't really talked much about how we track obsolete names of drugs and drugs in RxNorm. We have a fairly extensive process by which we follow closely. That even though the most minor change in a drug name for one of the sources results in our preserving the previous name as an obsolete name, and then if there is something an RxNorm primary name, like what we would call an SCD or an SBD, that's the RxNorm name for that product if there are no active names for that. Then we'd label that one as obsolete, but we'd still maintain it in the system because if we want to do historical records on various types of drug exposure issues or whatever. Then, you can at least know what was that drug at that time and what was in it, which can be very important. There are several child health epidemiology projects that are currently using RxNorm as their method of coding the drug exposers that they're following over long periods of time. So, they need that historical information available.
What are we going to do in the future is we're
going to take the RxNorm names, these names that we've created as the meanings of doing the interoperation between the various sources, and we're going to assign to them an NDC code as an attribute. And, what we're going to have to do is the editors are going to have to use some judgment in the cases of disagreement. I mean, there may be times when they have to say they can't decide, but we will have to look very carefully at each of those disagreements and see which we believe is most likely to be correct. We are also going to put into affect a procedure that I'm going to call cascading the listings. So for example, suppose we have a drug delivery device. It carries a branded drug, and this drug delivery device has an NDC code of 1. Then for the branded drug, we're going to say 1 represents also that branded drug in that sense, but 2 and 3 might be NDC codes; 2 and 3 might be just for that branded drug- various package sizes. And then at the generic name, we might have other NDC codes, say 4, 5, 6, 7, 8, 9. And we will add onto that the 1, 2, and 3 NDC codes. In that way, if you look to see if a drug has been administered-- if a prescription has been fulfilled by the particular NDC code. So, the pharmacy could dispense something with an NDC of 1 or 1, 2, and 3 and still it would be recognized as that prescription had been filled correctly. We're still going to maintain the source's assertions about NDC codes even if, and noting full well that there may be disagreements, so that one source may have an NDC code in one particular area and the NDC atoms themselves will carry the NDC code in a different area because we disagree with that source. But, it'll be our own assertion that this is where the NDC code truly belongs.
Well, another question is: What are we going to
do with the outdated NDC codes? What are we going to do when something— that package size is not longer available or that drug is no longer available. What are we going to do with that? And certainly with the RxNorm NDC code attribute when the RxNorm drug is obsolete, we can still carry those NDC code attributes. Beyond that, we do have some concerns about how to represent NDC code assertions from a particular source. One of the things I'm concerned about understanding a little bit better is: What the use case is for having these obsolete NDC codes? Because, if we're going to spend a lot of effort trying to make sure that we track them appropriately, we want to make sure the game was worth the candle. That it is significant. And that's something we're looking at. And of course, we have serious concerns about reusing codes. Practiced up until now and until the FDA takes control of the process, the practice has been that once an NDC code has not been used for five years, it may be reused. Now Joan Capusnik tells me that First Data Bank, which employes an army of people calling pharmaceutical manufacturers to track their NDC codes, notifies them if they start finding conflicts. I think that's absolutely fantastic, but we do need to get to a point where we can automatically recognize and prevent the reuse of those codes, and that's all I really have to say about all of this. I'm open for any questions.
MR. REYNOLDS: Okay, thank you Stuart. Excellent.
What I'm going to do, since we have two members and one staff member on the phone, I'm going to give you a first chance and a last chance. In other words, I'll let you ask your questions first and then I'll come around to you at the end also and it's not your last chance to answer, but…
DR. COHN: I have a question.
MR. REYNOLDS: Yes?
DR. COHN: Is Randy also testifying or not?
MR. REYNOLDS: Yes, as a second testifier at 10:30.
DR. COHN: Oh, okay. I feel like half the questions I have for Stuart are probably really for Randy.
MR. REYNOLDS: Well, should we have Randy go ahead and testify and then we'll ask all the questions?
DR. HUFF: I think it would be good to have Randy go because I think a lot of the questions are, you know, are there things that we can do so that everybody isn't trying to do the same work so that we can actually come up with an authoritative source for the NDC codes.
MR. REYNOLDS: Okay, so what we'll do is, we'll have Randy testify, then we'll take a break, then we'll come back and ask questions.
Agenda Item: FDA testimony—Status of NDC Database
DR. LEVIN: I was going to cover the following topics. First, the proposed changes to the FDA drug listing regulation that Stuart alluded to, and then discuss issues related to distribution of the NDC information, and then talk about the FDA medication information, identifiers, and terminologies that would be useable for electronic prescribing and some future initiatives that we're working on with Structure Product Labeling and drug listing.
First, I just wanted to note that the work that
we're doing is part of a broader collaboration that we're working on, and Stuart with the National Library of Medicine, but also with the Department of Veteran's Affairs and The Veterans Health Administration, the National Cancer Institute, and Arc on this federal medication terminologies. We're bringing to this activity our Structured Product Labeling with a variety of terminologies. And as Stuart was discussing, the source of new product information to our Structure Product Labelling, and National Library of Medicine, bringing in RxNorm in the clinical drug, and the process of linking this to the national drug code as well as the distribution of the Structure Product Labelling through DailyMed as Stuart was describing. The VA brings the National Drug File Reference Terminology with terminology for pharmacological class like mechanism of action, physiological effect, and structural class. And a problem with subset of SNOMED they created with Kaiser and they're linking RxNorm and the NDC with the other National Drug File Reference Terminology. The National Cancer Institute. The NCI thesaurus is a source for FDA terminology. We use the NCI enterprise vocabulary services to maintain as a source for our terminology and the source for a number of the terminologies we're talking about are available through the NCI Terminology Browser. And the Agency of Health Care Quality and Research with the work of Mike Fitzmaurice have been helping me with this project all along with various support and funding, so I just wanted to note that.
In August of this year, we published in the federal register a proposed rule to change the way we do drug listing. This rule has other aspects in it other than drug listing, but I know we've been talking about this for quite some time in this committee, and the proposed rule is finally proposed after all these years. But a couple things that this does is that it proposes that the FDA would manage the NDC, as Stuart was discussing, that we would be the central authority for the NDC. Right now, the manufacturer assigns the numbers. We provide a portion of the NDC, what's called the labeller code, the first part of the NDC, but the manufacturers then assign their own numbers. This would move the assignment to the FDA from the manufacturers. It would also, in this rule, lays out the specific rules. It clarifies the rules of what NDC means and when does an NDC change. So what characteristic of a product, what things have to change that would then lead to a change in the National Drug Code. So, this is in this proposed rule. It also automates the drug registration and drug listing process. It the moves it from what is now a manual paper based process to an automated electronic process. And one of the things we can do now that we have more experience with the Structure Product Labelling is to merge the issues related to labelling with the drug listing. And that's what we've done in Structure Product Labelling. We've taken the listing information. That's always been existing in labeling, but we've now put that in is as structure format in each label. And that's the structure product labeling that Stuart was describing that we distribute through with the National library of Medicine through DailyMed. So, this is a very important piece to merge these two activities together: Labeling and listing.
The other thing that the rule proposes is it
specifies which products can have a National Drug Code. Again, this is more of clarification. There have been products that have National Drug Codes that shouldn't have National Drug Codes, and this lays out what products can have National drug codes. The proposed rule takes into account a lot of the information we have received here at this committee as well as from other places so that we can better manage the National Drug Code, so we can meet some of the needs that we've been discussing over here in this committee as well as other places.
Now the rule was just proposed in August—end of
August. Then there will be comments to propose rule there. Other changes in this rule that may be fairly controversial for different groups that may, not necessarily related to things that are going on in this committee, but we'll have to handle these comments, and it depends on the number of comments received, but they'll be coming in over the next three or four months. Then, once the comments are received and addressed, and then it goes through that sign-off process that we all know about. Then, we can publish the final rule. The rule proposes that the industry has nine months to update their listing. After that time, the FDA takes over the listing, the management of the NDC. So they have nine months to make sure that their products are on the list. So, those are some of the changes that are being proposed. Again, this is a proposed rule, so we have to get the comments and make changes. But again a lot of the things that are in this rule are things that were discussed here and other venues about what's necessary for the management of the National Drug Code.
Now because this proposed rule, what this would
lead, is that the FDA National Drug Code Directory would contain the definitive list of valid codes. Right now that is not so. And as Stuart has brought up, there are a number of problems with National drug code directory. But because we would be assigning the codes, we would know all the codes. Right now, that is not true, so that is one thing that would change with this regulation. Another thing is that this Structure Product Labelling could be then used as the method for electronically distributing this information at the product level. So, it could be up-to-date information, as Stuart has noted that this would be a place then that they could go to get the information, the National Drug Code, at the product level. The Structure Product Labelling contains not only these drug listing data elements, which I'll go through in a minute, but it also contains all the contents of the labelling and how to use the drug and all the prescribing information and other information about the drug.
Now, as far as the status on the structure
product labelling-- by the end of this year, we expect to have the Structure Product Labelling for the approved prescription on innovators. This is not the generic manufacturers, but the least the innovators on the DailyMed web site. Then, over the next few years, then we would be going into a maintenance process. Again this is depending on the rule, when it goes into affect. We have maintained the up-to-date Structure Product Labeling for marketed drug products, both human and animal drug products, and maintain on the DailyMed web site. So that would be the method for distributing the National Drug Code information.
Now the drug listing information, that's in the
Structure Product Labeling, describes the product, has the proprietary name and a nonproprietary name (The generic name or the brand name, however you want to describe that.), and the code or the product which is part of the National Drug Code now. The National Drug Code of course has the three pieces. The labeler talks about who the distributer is, the manufacturer distributed the product, code talks more about the product, and then there's a package component. If you look at the first two components, you have a code that would be specific for product. Now this again, because of the way that the manufacturer assigns the codes, we can't be assured that this is correct. But with the way that the proposed rule has talked about management of the NDC, this would happen: There would be a code specific to a product as defined in the rule. So the rule will define what a product is and there would be a code assigned to it and if anything changes in that product, a new code gets assigned.
We have in the drug listing information, also,
the ingredients, the active and inactive ingredients, and the active moiety. Each of these has a name and a code. The code that we use for ingredients is the unique ingredient identifier from the FDA Substance Registration System. So, there we would have for each of the active ingredients and the active moiety there will be codes. And I'll go into more detail on this in a little bit. There is also the active and inactive ingredients strength. That would be that we use the codes that we've worked with with the National Cancer Institute and the UCUM codes for strength. Dosage form is also in the drug listing information. The appearance of the product is also in the drug listing information, like anything that's printed on the solid dosage form, the color, shape, size, scoring, coding, symbols. Those are all part of the drug listing information. Route of administration is part of the drug listing information as well as the DEA schedule. And for packaging, it's the package type and the quantity and code for that. And the code for that would be, again, the National Drug Code, which includes the packaging component. All these things are defined in the proposed rule and then would be managed by the FDA.
This picture just summarizes identifiers. There's
an active moiety, again the unique ingredient identifier from the FDA Substance Registration System, this is part of the active ingredients. There may be more than one active ingredient in a product. If you combine all the active ingredients with the strength, the dosage form, the inactive ingredients, and the appearance as well as the manufacturer, you get a drug product, and you get a code associated with that. There can be more than one drug product in a package combined, and Stuart mentions something along those lines, and then the package type and quantity leads you to the packaged drug product, which is what the National Drug Code is associated with. Again, all of this, the rules would be defined in the regulations and managed by the FDA.
I'll just go through some of the FDA medication
identifiers: the ingredients, the products, and the package products. Ingredients: There is a substance that makes up the product. There is an active moiety, which is the active ingredient without the counter ion. So, an example of ingredient is oxycodone hydrochloride, the active moiety would be oxycodone. And then, I just provide some ingredient identifiers. We have an identifier for oxycodone hydrochloride. We have an identifier for oxycodone. There's a separate identifier for acetaminophen. We have identifiers also for foods. This is part of our activity for the food allergy activity. So, you know, we have an identifier for things like peanuts or milk as well as identifiers for botanicals. So this year, we expect to have the identifiers for all the active ingredients in the approved prescription drugs. We have identifiers for all the food ingredients that came from the VA food allergy list and most of the botanicals identifiers sends for some inactive ingredients. After the rule goes into affect, we'll be maintaining these identifiers for active ingredients for other products such as over the counter products, biologics, and animal drug products. And we'll also, over time, be adding more inactive ingredients for marketing products.
For the drug product, and again there is a
definition for the drug product now, it would be in the proposed rule. An example of that would be Tylox. It's the proprietary name for the product. A nonproprietary name would be oxycodone hydrochloride, and Acetaminophen would be the generic name. There is the product code that is associated with Tylox and shows there is the product code. That's defined by the characteristics of that. Again, that's in the proposed rule, what exactly is a drug product and what would change. Right now, we have most of the prescription drugs in the National Drug Code, and as Stuart noted, there are some problems with the National Drug Code. And by the end of the year, we're looking at the SPL with the drug product codes for the approved prescription innovator drugs. Once the rules goes into affect, then we will be able to maintain up-to-date Structure Product Labeling with the drug product codes, would be in the Structure Product Labeling for marketed products. Again, that's with human and animal products. So you have Tylox who is the example of a drug product. Now Tylox and 100 capsules in a bottle would be the packaged drug product, or Tylox 100 capsules in a blister pack would be a packer's product. Each one of those would have their own National Drug Code. And that status of the packaged drug product is the same as the status of the drug product.
We have medication terminology that we maintain. The manufactured dosage form, the route of administration used to measure appearance and package type. So, all these are available in the NCI Thesaurus. We've been using these codes for quite some time in our drug listing process. The manufactured dosage form examples would be a tablet, orally disintegrating; powder, for suspension; injection, for solution. These each would have its own code and be available in the NCI thesaurus.
The route of administration, examples like oral,
intramuscular, intrathecal, these are all in the NCI thesaurus and currently available again being that we've been using those for a number of years.
Unit of measure: We use the Unified Code for
Units of Measure (UCUM) as well as some translations of units of measure that we supply. So for example, a unit of measure might be milligrams per tablet. The milligrams (mg) is the UCUM code. We have a code or translation for that as just spelling it out as milligrams. But the per tablet that this unit of oneness, this tablet, is not part of the UCUM codes that we provide the codes for those, and that this would relate to so this is strength would also the same thing used for product quality. So if you say there are 30 tablets in a bottle, the tablet there is not the dosage form but this unit of measure for how many milligrams in this unit. It could be milliliters per spray. There are a variety of types of units of measure so that the code for spray comes from the FDA that's in the NCI thesaurus. All these are available in the NCI thesaurus and UCUM is also available.
Appearance is for solid, oral, dosage forms.
These are the color, shape, size, coating, scoring, symbol, imprint code. We have codes for colors and shapes. They're not just text fields buts specific codes. And the appearance would be used when we receive calls, or poison control centers receive calls that they see they get this drug from the pharmacist, and they want to know if it's the drug they're supposed to be using. And they'll give the appearance of it, and we would have the data in our Structure Product Labeling so you could find out what that product is from the appearance.
Package type: Things like a bottle or a dialpack.
Those are again in the NCI thesaurus for any of the package types. Some other initiatives, some potential future initiatives for the Structure Product Labeling would be possibly adding additional data elements. One are the data elements called Highlights Data Elements which I'll go over in a minute, but another activity would be adding something like the NCPDP billing Units. So NCPDP has come up with a standard way for pharmacists to build for the products. And we've been discussing with the NCPDP that that billing unit could be added as one of the units in the structure product labeling, provided by the manufacturers so that way it's in the product labeling of how you would bill. So their billing units are such as thirty each. Like Tow(?) would say 30 each or something. However they describe their billing would be added into the Structure Product Labeling as a way to transport this information from the manufacturer to the prescriber. This billing unit is not really an FDA activity, but it would be using the Structure Product Labeling as a way to communicate between the manufacturer and the health care community with some of this information. So, billing units is one example, but other information can be put into the Structure Product Labeling as a way to communicate this kind of information. Then would be available on DailyMed for download and use by information suppliers whomever they would be.
We've also been looking at the Structure Product
Labeling for other products that the FDA regulates, for example medical devices. There's been a lot of discussion on medical devices and how to define what a medical device is. The Unique Device Identifiers. We're going to have a public meeting on that soon. One of the concepts we've been talking about is just as we have a drug listing component in the Structure Product Labeling, we could have a device listing component as well and describe what a device is, what components change in the device that would lead to a new code. So, this is something that we're working on.
Dietary supplements, Structure Product Labeling
could be used for dietary supplements for food products for other products as well. One of the dietary supplements are one of the components that are not drugs according to the FDA regulations, and they do not receive National Drug Codes. Though some dietary supplements today have a National Drug Code associated with them by the manufacturer. But this is one of the things that if the FDA is managing those products, that this would not have a code. Whether it's important for them to also have a code from a central authority or whatever, that's something we would need to hear if that's something that's important.
I'm just going to finish up going over some of the Highlights Data Elements. This is one of the new initiatives for the Structure Product Labeling that we're looking to add these codes. One of these is for indication, including the intent of use. And the indication that this is a treatment for migraine, or this is a prophylaxis for stroke. That would be coding that information in for usage of the product, the maximum dose, so many milligrams per day, that kind of coding. For condition of use is just an adjunct, like it should be with aspirin. We would code that information in. If there are certain screening tests that need to be performed, or a person should be monitored, we would code those tests in like they should be monitoring the absolute neutrophil count or something along those lines. Also, if there's a condition of use, it should only be used in a certain population, patient characteristic, use in children under the age of six could be coded in. These are in the label as a narrative text, but this would be data elements that are machine readable following control terminology that would be coded into the system. Also, coding in some limitations of use like contraindications or when you should use something with caution or modify the dose where there's little or no information, or if it's a second line treatment. There is some issue about the duration of use, or that we approved it on a conditional basis. Those types of information, like a drug is contraindicating women. We would code that kind of information. Or the dose should be modified in patients with renal impairments. Again, coding in that information. More interaction on Highlights Data Elements would be interactions such as we would code a contributing factor and consequence. Use with Dilantin has a consequence of increasing drug levels. That kind of thing would be coded. Again, these are all using control terminology in machine-readable codes. Use with MAO Inhibitors can cause malignant hypertension. Coding adverse reactions could be coded as well as the pharmacological class of the mechanism of action, physiological effect, and/or structural class as codes.
A lot of the references for Structure Product
Labeling and the terminology are on the FDA data standards council web site at www.fda.gov/oc/datacouncil and you can see information about structure product labeling as well as other standards that we're working on at FDA. Thank you.
MR. REYNOLDS: Thank you, Randy. Nice job. What I will do is go in the order of Justine, Stan, Jorge, and start making the list in the room. And then at the end, I'll come back around to catch you guys again so that you get equal opportunity. So Justine do you have any questions?
DR. CARR: Yes, I do. Thanks. Both very interesting presentations. I wanted to understand a little bit more about what we were just hearing about of the coding into the label, the adverse reactions, or the interactions and understanding how in terms of electronic prescribing, is the intent that that code in the label would then alert providers about drug interaction? I guess, perhaps, if I step back. I'm trying to understand where on the continuum of drug prescribing this information would add value, and also would it be limited to the information available at the time drug is approved, or is there some mechanism for updating information as it becomes available after the drug is approved?
DR. LEVIN: Well, I'll talk about the procedure. The idea would be that this would be conveyed information that's in the label. So yes it would convey that information. But as the labelling text changes, that this data elements would also change, would be updated. So if there is a new adverse reaction and it finds itself into the label, and then that could also then find its way into the codes. So, the information is to convey information that's in the label. The thing that we've been trying to do is, with the Structure Product Labeling, is get the information in a machine readable format and then make it available in a standardized format like the Structure Product Labeling distributed through the National Library of Medicine DailyMed site, and then let people take that information and use it as they see how it could benefit. We weren't necessarily defining how it would be used, just that we would make it available for use.
DR. CARR: Thanks.
MR. REYNOLDS: Stan?
DR. HUFF: So I know you said, Randy, in your testimony what the timeline was for the proposed rule, but could you just repeat that? I didn't catch that.
DR. LEVINE: The rule was proposed at the end of August. Then there is a comment period, and it's hard to know the length of the comment period because people can request extensions for the comment because it's a fairly extensive rule. So the comment period may go for three or four months lets say. Then, once the comments are received, the comments have to be addressed. Once they are addressed, then they need to be agreed upon, not just inside the FDA but at the department level, and then the office of management budget looks at the final rule, and then the rule will go final. Once the rule goes final, and the proposed rule, there is a period of time where the manufacturers would be allowed to make sure that all their products are currently on the list of valid products. So I think that right now, the proposal is, I think, a nine-month period after the rule goes final. So they have that period of time. Once that period of time finishes, then the FDA takes over the management of the National Drug Code. That means, the reason for this is that every National Drug Code currently in existence could be maintained, could be continued to be used. Instead of doing just a broad change of all the National Drug Codes, but that requires you to make sure the information is available to the FDA so that it can be on the list, but once that period of time finishes, then if you want a new NDC, you would need to go through the process to get the NDC through the FDA.
DR. HUFF: So it sounds like it might be easily a year and a half or two years before you're actually making assignments and covering that.
DR. LEVINE: I don't know the exact timing, but you know how we've been trying to work on this propose rule.
DR. HUFF: Right.
DR. LEVINE: Okay, so that gives you just a clue. If there are circumstances that warn people moving things along quicker. I mean it's just the way the system works.
DR. HUFF: Sure. Well the comment I should have made first was just to express my appreciation for all the work you're doing. How you and the National Library of Medicine, these are wonderful useful things that are really hard to do. And I know they're really hard to do, especially when you need to worry about the proper timing and careful management of the regulation process, but you know it's really wonderful to see the development of these identifiers, which are really going to establish a computable basis for medication use and understanding of how we're using them, what the interactions are. It's really wonderful work. Thank you for that. It's unfortunate it takes so long, but I realize it does. But we really appreciate what you're doing.
My other question was for Stuart. As I understand, Stuart, right now the NDC codes are listed as attributes associated with the generic or the RxNorm clinical drug sorts of things. In our own electronic medical record data dictionary, we originally had the NDC code as essentially synonyms or alternative representations for generic drugs, and then after awhile we ran into trouble down that path and said, ya know, we hate to do it but we're actually going to make these 300,000+ codes concepts because that's what they really represent. They represent something that's essentially a drug product associated with specific packaging and so there's defining characteristics for those things that allow them to be a concept and then we treated the rest of the things as relationships to say that there's a relationship between the generic drug and this NDC level concept which I think a package product might be the right name for what the NDC Code represents. And I wonder if you guys have been thinking about that and what you see as pros and cons of making the NDC Codes actual concepts in the Medi Thesaurus rather than as attributes on the other concepts.
DR. NELSON: Thank you, Stan. Our primary difficulty in putting them in as concepts in the Medi Thesaurus is always revolved around the problem of tracking the meaning of those NDC codes and obtaining the meaning of the NDC codes. It has indeed been somewhat problematic and remains somewhat problematic which is why we're very much welcoming the FDA's proposed rule about controlling the NDC Codes. I would wonder what additional value you would obtain by making them concepts as opposed to keeping them as attributes of the particular products. I'm not sure of the added value in doing that. I certainly would be open to hearing more about that, but at one point we discussed whether or not we were going to try and track NDC Codes as concepts in the Medi Thesaurus and decided that it was not worth the effort to try and do that and maintain the history of concepts that we try to maintain in the Medi Thesaurus and so forth. Certainly have an open mind about the issue, but I would like to know... Ya know, we don't treat them as synonyms. We treat them as attributes. We say that they are codes that represent specific entities that we are representing at the RxNorm level. RxNorm level is a specific level of abstraction about drugs. These are the things that fit within that category of that abstraction. I think that that is all we can say right now with that. I'm not sure. I guess the question back to you would be: What added value would I get by making it a concept?
DR. HUFF: I need to think about it and make sure that it is worth it. What you get is, you actually have a defined concept and there are more questions you can answer than if you know that this NDC Code was given, and it does in fact have a precise computable value that you can determine from the database so that you know the manufacturer, you can imply in fact the exact drug product that's in there. Then there are more things that I know about how this thing was given and prescribed than if its--- but all of that value turns on exactly what you said. And it's probably premature to do anything now, but in the future situation where we have well defined and controlled rules and processes about when it's created and the meaning of the NDC Code in terms of exactly how it's used and we're not assigning them to things that shouldn't have them, and we're not assigning them to devices, but taking care of all those issues. Then I think there are things you can know about the prescribing process and about he association of how often this packaged drug was given versus when that other packaged drug was given in tracking of potential with the supposed non-ingredients of the thing, etc. that would have value. But I need to think about it and make sure that there is value there.
DR. NELSON: Alright, well I'll look forward to hearing from you one way or another. You know where to find me.
DR. HUFF: Okay. That's all I had for now.
MR. REYNOLDS: Jorge, did you have anything?
DR. FERRER: I don't have any questions.
MR. REYNOLDS: Okay, it would Simon and then Mike and then Wanda. And what I'd like to make sure that we do, is looking up at the clock, we got about twelve more minutes for Stuart, which I actually have one. But as we go through this let's make sure we honor his time and make sure we get our questions to him. Because Randy was supposed to go to 11:30 so we'll make him stay that long and talk to us.
DR. COHN: Well I'll ask a couple questions. I get a little confused about who, and this is also tightly intertwined, as it should be, that it's hard to know who to look at with some of these questions. Let me just start with one which is, I haven't read the proposed rule. IS RxNorm is the actual code on the box? Is it going to be printed there?
DR. NELSON: No.
DR. COHN: Why not?
DR. NELSON: Well, I think that the RxNorm name and code, and by code here I believe we're talking about the concept identifier. Identify is a whole number of different names for product, added given level of abstraction. Now a generic drug might represent at that level of abstraction six or seven different things that the FDA would call a product, meaning they have different manufacturers or whatever. So, that at that level of abstraction, it's quite different.
DR. COHN: I know all that. But it is the same medication, right?
DR. NELSON: Well…
DR. COHN: Or is it not the same medication?
DR. NELSON: Simon, this is the tricky part. It's what makes it the same medication. You know, when are two things identical? It's a deep question. I would put forth to you that every manufacturer would like to say that their products are unique. To some degree that may be true. But we're saying here, under the generic drugs, because perhaps the excipients are different so that it depends upon how you look at it whether or not they are identical. You can say that they belong to a given category and that category as expressed by the generic active ingredients and strength, they all belong to that category, but they're not necessarily identical.
DR. COHN: Okay. Let me explain the reason I'm asking. If anything, most of us see RxNorm as being helpful in the actual prescribing of something that has an NDC Code. At least that's what I think most of us think that it has a place there. And so I've heard from many people that it would be very valuable if we were uninreusley(?) knew for an NDC Code easily what the RxNorm code was so you could put it into a prescribing system and knowing that you were getting if not one NDC, ya know, there's always that issue…
DR. NELSON: Exactly.
DR. COHN: And the amoxicillins that you get, and they all have different NDC Codes, and…
DR. NELSON: And that's why we carry it with the RxNorm names so that you know out of the 287 ampicillins 250 milligram tablets, they have 287 different NDC Codes for that, that you know that every one of those NDC Codes fulfills that particular prescription.
DR. COHN: Well, Okay, and so that's in the database, it just isn't on the container?
DR. NELSON: It's in the way that RxNorm represents the NDC Codes as attributes of that particular code. So we'll say for ampicillin 250 milligram oral tablet, these are the NDC Codes that fulfil that requirement at that level of abstraction. And what I'm saying is, at that level of abstraction is a level of abstraction, at least in my experience, is a level at which practitioners talk.
DR. COHN: I think you're doing everything but answering my question, but I do appreciate… My question is should it be on the label or actually on the box?
DR. NELSON: I'm saying it's not. Now I will tell you that at the DailyMed site, after we get the product structure label we are going to associate that product label with an RxNorm name. But understand that there might be six different labels that are associated with the same RxNorm name.
DR. COHN: That's my point. You got it. Thank you. That was my point. Now, Randy, why isn't it in the proposed rule and you serve a question. I mean, do you have a thought on that?
DR. LEVINE: In the proposed rule is a requirement to put the NDC on the package. That has not been a requirement up until now, but it is being proposed. And then the NDC would be, when the label is being put together for a new product, we won't know what the RxNorm code is, because it hasn't been created necessarily yet. And so it's only until we send it to the SPL to the National Library of Medicine. Well, they assign at the RxNorm code. I mean, Stuart, am I getting that right?
DR. NELSON: That's correct. We don't have access to the FDA's proprietary information before a drug is approved. Before a drug is approved, we would run as approved the label is released. Then we get it, we can make an RxNorm name for it. Until that, we don't have access to that data. So it may be a process. It may be something we can work out about having some kind of advanced notice with an embargo or whatever. But until we have that data in advance, we can't give it an RxNorm name. Right? And an RxNorm identifier.
DR. COHN: So that actually even includes such things as changing bottles from 30 to 25, making a new tablet green rather than yellow and all of that which should get you new NDC Codes. That correct?
DR. LEVIN: When you make those changes, then you get a new NDC Code.
DR. COHN: And basically you're bargoing(?) that from NOM until such time as it is publically released with that new container and new bottle, even if it's Amocillin 250?
DR. LEVIN: Right. Once it's made public, that's when..
DR. COHN: And I apologize, I'm just trying to figure where all of this fits. Exactly how valuable and useful, and once again as a patient's safety concern and other things, this issue of putting the RxNorm on the bottle would be a very helpful thing or on the container. And so I was just, I guess I was hoping that that would be there and that there would be obviously tighter coordination between FDA and NLM in terms of all of this sort of happening together as opposed to sequentially. So, it obviously speaks the fact 80% of the way there, I'm obviously speaking about the state. Where we're really bringing patient's safety first, I guess is sort of where I was trying to go.
DR. LEVINE: We proposed the rule, and there is a comment period.
DR. FITZMAURICE: Thank you. Now, one other question just to make sure I understand how things are identified in all this stuff. I heard, obviously, that NDC is being used in many interesting ways. And so my understanding is in the future state. Yes, there will be a 100 pack of something or other, of Amocillin or whatever and that will be in somebody's pharmacy or, say, in a hospital. And then they actually dispense it and give it to the patient. Well, at that point it sort of transformed into an RxNorm dispense code because there's really no way to represent NDC at that particular level of annularity. Is that correct? Is that how it sort of plays as we move here?
DR. NELSON: Well I would think that would be one way could play. Another way could play would be that the NDCs do have three parts to the NDC code. And the first two parts together constitute an identifier for the product, which specifies the manufacturer. It doesn't specify the package size, but it is for a given product. That would be one way. I'm kind of interested, maybe could sneak in a question to Randy, because this is very closely related. And that is, Randy, you talk about NDC codes being in the SPL but you're not going to issue a new label every time they have a new package size, so what you're really talking about is having the drug product code in the SPL. Am I correct?
DR. LEVIN: No, every time there is a new package size you get a new label.
DR. NELSON: Really?
DR. LEVIN: Yep.
DR. NELSON: Oh boy, we're going to be working.
DR. LEVIN: Because in the label is the how it's supplied. So if a company comes up with a new packaging, they can update their label with that information. They definitely update their listing information with that. So once they provide new listing information. So these are two components that have in the past been separated labeling and listing. But if something changes with the listing, that's part of the structure product labeling now.
DR. FITZMAURICE: Randy had a nice listing of the partners that the FDA had to do its portion of the Structure Product Labeling and shipping the information over to NLM. Does NLM also have a list of partners?
DR. NELSON: Well, I would be remiss if I did not acknowledge the key role that HRQ has played in the development of all of this. I would be very remised not to because they certainly have been a very valuable partner. And clearly we've tried to work closely with the FDA to make the DailyMed work and to obtain that information and try to understand where we stand with things like the FDA substance registry system and what those codes are and make sure that we get those involved at the ingredient level and so forth. And working with the VA and the Department of Defense as well to try to help them do what is really a massive interoperation effort that the DOD and the VA is trying to do. So I would have to acknowledge all of those folks as very close partners. I think the NCI has helped a great deal with the terminology services for the FDA that the NLN has not really been able to provide for them. I would like to acknowledge that as well. You know, I don't think any of it could have happened without HRQ and their support.
DR. FITZMAURICE: Thank you, and I would say that support is an effort that not only NLM and FDA put together, but also the office of the assistant secretary for planning and evaluation. There's been a coordinated effort to get funding to this very important patient safety endeavor. And also, I would like to respond that the Electronic Prescription Pilots didn't come to NLM for advice. If we haven't, we should have. If congress should mandate additional pilots or if issues should come up in the current pilots, we will be knocking at your door. Semantics are an important part. I think many of them used NDC Codes because that's what was on the package and related to what the doctors were prescribing. But it would have been a good pilot if we had more time and more money to use RxNorm's and link it with the drug information. Some of them have tried to use RxNorm.
DR. NELSON: Yeah, I think it would be. Also, I'll tell you frankly, the last two years in RxNorm business have been a roller coaster because it seemed to be an idea that took off before we actually had the wings on the plane. We've been producing a new release of RxNorm monthly while implementation and further refinements of the system that maintain the vocabulary are still going forward. And so it's been a wild ride. But I'm happy to say that we've had a good team, and we think we have high quality information and I think we've got a good product coming out. So, we're working as fast as we can and as hard as we can.
DR. FITZMAURICE: I want to follow up on what Simon said, because Simon raised a really important issue. Is there anything that stands in the way of NLM shipping of their information on a new drug or new package or new label to NLM. How long does it take NLM to assign an RxNorm and then get it back to FDA?
DR. NELSON: What I am advocating and pushing my team to be able to do, is to have it within a week. To at least have the preliminary identifier on it and the attributes and that means we've received it, we've processed it, and a human has sat down and looked at it and said, “Yes, it fulfills our rules.” We have very strict business rules about how assign strengths for medication, fordoes forms, and so forth, and a human has to really look and make sure that all of those things have been done correctly. So, I would say that my goal is within a week of having to receive the information from the FDA, we will turn around and have it ready.
DR. FITZMAURICE: And the preliminary fractions needn't add a week to the process. It could be done sometime before FDA issues the…
DR. NELSON: If we did, I understand the FDA would have concerns, and I would have concerns as well about making sure that we don't release proprietary information before its release date. I work at a library. We make information available to people. We're not in the business of hiding information. So it kind of takes a little bit of a different mindset about how to do it. I think it would be possible to do that if it were only internal in our system to do that. That's probably something Randy and I need to… Randy had lunch a few years ago at an HL-7 meeting. And Randy said, “Would you guys be interested in having labels?” and I practically fell out of my seat. Maybe Randy and I need to have another lunch where we could talk about futures.
DR. FITZMAURICE: Stuart, slide number 6, you talk about other problems with other NDC codes and you say there is no semantic link. Do you mean that a drug might have different names and that there is no semantic link to the other names? I'm not quite sure what that meant.
DR. NELSON: There is no way of knowing short of something like RxNorm or another source that you could find out what NDC codes are for the same types of products.
DR. FITZMAURICE: So there's no good classification?
DR. NELSON: It's not a classification. It's an assignment of an identifier. Yeah. For better or for worse, I think that the NDCs were developed a long time ago, and I think at the time it was very farsighted for them to do so. I think now the time has come to look at it and say, “Yes, we to get the horse back in the barn.” Let the FDA control those NDC Codes. Let's make it a little bit easier for people to have access to the information of what is involved with them.
DR. FITZMAURICE: My last question is: Is the same information in both NLM's UMLS Medi Thesaurus and NCI's Medi Thesaurus? I know that there is a lot of mapping of RxNorm to NLM's Medi Thesaurus and that a lot of the semantics and the units come from NCI's Medi Thesaurus. Is there good partnership? Is there a reason this should all be in one place, or can things be electronically so it doesn't hinder anybody?
DR. NELSON: Well, I think the NCI has had a somewhat different model with what they want to do with their terminology services. For example, what they've done with the subsets for the FDA terminology, they've done with lots of other kinds of terminologies specifically related to cancer. It all gets incorporated into the ULMS on a periodic basis. I think it would be better for us at RxNorm to be able to have direct access to those FDA medication terminologies subsets, rather than try to go through the NCI Thesaurus because then you have to wade through a whole lot of material to extract a small amount.
DR. FITZMAURICE: That raises a business use question. Are both available for private sector production work or available just for research? Is it something of like a first data bank, a multum could access these files and use them to add information to their products?
DR. NELSON: Well, the ULMS has a fairly extensive licensing procedure because we have proprietary naming products in the ULMS as well as things that are public domain. Now, for example, in RxNorm, is first data bank is proprietary. If you're going to use their terminology and their codes, you need to have a license from first data bank. On the other hand, the RxNorm names and codes are right there right next to it, and they are not proprietary. Okay? So that if I don't have a license to use First Data Bank, I can still use the RxNorm portion of it.
DR. FITZMAURICE: Can you speak to the National Cancer Institute or is that for somebody else…
DR. NELSON: I think that the National Cancer Institute would have to tell you what vocabularies they have that are proprietary are and non-proprietary. And of course, all of us, we have a National license for SNOMED, so anyone in the United States can use SNOMED which is proprietary but essentially a national license so they can use it.
MR. REYNOLDS: It's past your time, can you take a couple more, or are you out of here?
DR. NELSON: Sure, we'll just keep our fingers crossed that my meter doesn't run out.
MR. REYNOLDS: Okay. Did you have one for Stuart?
MS. GOVAN-JENKINS: This is my question: Prior to FDA, how and who were the NDC codes generated in? I'm just kind of concerned about duplication. Maybe I don't know the history behind it, but duplication of NDC Codes as far as quantity control?
DR. LEVIN: The NDC codes were, the system was set up many, many years ago and the way that it works is that the FDA provides, it's the three codes, three parts. The first part is what's called the labeler. And the FDA assigns that to manufacturers and it will repeat. We only ever assign that once. I'll give you a labeler code then you go off and assign the rest of the NDC for your products. So you're going to come up with a new product, you're going to give it a new code using that labeler code. That prevents duplication because you own that domain of NDCs so that you can assign that. Now, that said, once you take the hyphens out, and you just have a string of numbers, there can be duplication because the way that the National Drug Company works is the pattern of the ten digits can change as there are only three acceptable patterns and they can be duplicated if you remove the hyphens. There is that potential. Usually because this would cause problems in the market, you would probably find that out pretty soon that if you duplicate it. We look through our codes. We found one duplication. Theoretically it could be a problem. But you own that labeler code, and that way that helps you prevent from duplication. With the FDA managing it, then there are stricter rules, including that we wouldn't have any duplications.
MR. REYNOLDS: Okay, I got one and then Karen, and then we'll see if we can get back to ya. And mine's just a point. You don't look at your two presentations because Randy on the package drug product, you list a number that's 12 including dashes, and you were talking about 10 and 11.
DR. REYNOLDS: Without hyphens. You can take a ten-digit, hyphenated, and basically of the three acceptable patterns by padding with a zero you can essentially make a unique eleven-digit identifier out of the ten digits.
MR. REYNOLDS: No, I understand. But if you look at…
DR. REYNOLDS: It's twelve digits if you include the hyphens, but you know, so if I say it's a twelve-digit versus an eleven-digit without hyphens. If I put in a zero in the right place, then I can pad it. I can make it a unique number. And the fact is that various sources, various pharmacy knowledge base vendors, may or may not do that and may or may not indicate where they padded that extra zero. Okay?
MR. REYNOLDS: Okay. No, that's fine. I heard ten and eleven, and I saw twelve and I just wanted to at least make sure…
DR. LEVIN: The NDC is ten-digit. The dashes are not. We don't include those as a digit in the NDC. Because in the barcode, you don't represent those dashes as a series, and that where is you can potentially have duplications because some of the times the labeler code is four digits, sometimes the labeler code is five digits. Sometimes the product code is three digits; sometimes it's four digits. So there are three different patterns. And if you remove those dashes, you don't know which pattern, just by looking at it, you don't know which pattern it is.
MR. REYNOLDS: Okay.
MS. TRUDEL: Yeah, Stuart, a quick question. You said in your presentation that currently one of the issues is that no one source is authoritative.
DR. NELSON: No one source for the…?
MS. TRUDEL: For the NDC.
DR. NELSON: …for the NDC is authoritative? Yes.
MS. TRUDEL: So the rest of my question is then after Randy's final rule goes into affect and nine months later, then the FDA takes over the maintenance of the NDCs. Does that become the one authoritative source for National Drug Codes?
DR. NELSON: I sure would hope so.
MS. TRUDEL: Okay.
DR. NELSON: And I'm looking forward to that day.
MR. REYNOLDS: Okay we have a question from the audience if you could push the button on that and introduce yourself and either make your comment or ask your question please.
MS. CAPUZNIC: Hi, yes, my name is Joan Capuznic. Stuart referred to me in his comment previously. We so submit our information on a regular bases to the NLM. And again, that facilitates the creation of the RxNorm concepts. Both happen after the fact after the NDCs released from the FDA, and there is a huge timing issue. There is also the research that has to occur to say, “Is this a unique concept or not?” And NDC that's released by a manufacturer and they hope for uniqueness in the marketplace, of course. We have to decide: Is this really a dose form that we already know about, is it a strength that's reasonably new? They might change something by a half a milligram to be unique. So there are a lot of issues regarding that RxNorm concept identification. So there's a lot of research and work that goes into that.
Also, as far as the NDC is a lot of information has flowed about the patterns of use in certain business requirements, in certain forms that are filled out, there is a specified number of byte spaces that need to be used, so that the filling zero is actually important for certain business purposes, and it would be great to have rules around collapsing and assigning those things so that we know that if a zero has been filled out for one purpose, that you can actually collapse back and find out what NDC was actually prescribed in another system that there is synonymy known there.
MR. REYNOLDS: Okay, any other last questions for Stuart so we don't his car towed? Stuart, thank you.
DR. NELSON: Thank you.
MR. REYNOLDS: I really appreciate you coming and joining us. What I'd like to do is, let's take a break until 11:10, and we'll be back. And then Randy, we can continue with the questions for you. Thank you.
BREAK
MR. REYNOLDS: Justine and Stan and Jorge are you back on? Okay, everybody back on the phone? Justine?
DR. CARR: Yes, I'm back. I'm going to have to get on off five minutes.
MR. REYNOLDS: Stan?
DR. HUFF: Yes, I'm back.
MR. REYNOLDS: Jorge?
DR. FERRER: Back.
MR. REYNOLDS: Okay, good. Alright, we'll continue our questioning of Randy. So let's go back through it again. Justine, do you have anything else for Randy?
DR. CARR: Not at this time, thanks.
MR. REYNOLDS: Stan?
DR. HUFF: No.
MR. REYNOLDS: Jorge?
DR. FERRER: Not at this moment.
MR. REYNOLDS: Okay, Jeffry?
MR. BLAIR: Not right now.
MR. REYNOLDS: Karen?
MS. TRUDEL: I do have one question. What I have heard from a variety of sources is that the NDC directory currently does have obsolete NDCs in it, but there may be NDCs that are not there at all, that actually belong to real live products that are on the market. Of course, they're not linked up to RxNorm. So, my question is then: at what point in time will we have a freely available, noncommercial directory that will have the complete up-to-date list of codes and that those codes will be associated with the appropriate RxNorm codes. And I think that what you're saying is you need final rule, plus the nine months to take over the database, plus the period of time (like 2007-2010) to link it up to the structured product label, and at that point we would have a complete authoritative database of NDCs that are associated with the proper RxNorm concepts. Is that right?
DR. LEVIN: Those are only estimates, but everything but the last piece, the hookup to the RxNorm. That is when we send it to the National Library of Medicine. But you're right, the process would require us to be able to move this information from our listing to SPL and then they would be able to provide the link to the RxNorm.
MS. TRUDEL: Okay, and then I'm just looking at the time table here: 2007-2010 would be needed to completely get all of the drug products into the SPL. So that they would all be there in that place. 2010 about?
DR. LEVIN: It's as much as I can guess based on what has to happen with the rule and it will take time to get that, but it will be something that can happen if the rule goes into affect.
MS. TRUDEL: Right. Great, thanks.
MR. REYNOLDS: JEFFRY?
MR. BLAIR: Randy, the presentation you gave was very very helpful and I realize there's things in the Structure Product Label that I was not aware of before. You mentioned that some of the elements that are in there, even coded elements, include the ingredients for a drug and to what degree are the codified ingredients overlapping or redundant with RxNorm codes?
DR. LEVIN: Well, the way I understand it is that the National library of medicine, they take the information that would be in a Structure Product Labeling which would be the ingredients and then use those to create the RxNorm code, like a clinical drug code, but they would create that based on what they see in the labeling, but they do something in addition because one of the things they have is the clinical drug is based on the order-able dosage form and what we have is the manufacturer's dosage form. So we will say it's a powder for suspension. That's how we would classify the dosage form. But they, I think, would say is a suspension. So that's a difference. But I think they could use the information that's in the structure product labeling to use that to create the RxNorm.
MR. BLAIR: Well, then let me ask the question a little differently. I thought the purpose for why RxNorm was being created was to give the clinical drug ingredient codes defined in such a way that it would be helpful especially any prescribing for drug-to-drug interactions, drug lab, drug allergy, all the rest with some clinical specificity. If FDA is now providing that in the Structure Product Label, why do we need RxNorm codes?
DR. LEVIN: In the Structure Product Label, we don't have the elective code for the grouping of all the ingredients, the dosage form and the strength, we don't provide a code for all that group of entities. Though, if it was automated, we could provide that, but we have not been and that's not one of the things that's been given to us as like it's a need or that's what we should be doing. So what we've done is, we've created the codes per the ingredients, the dosage form, the route of administration, and then handed that over to National library of Medicine that create RxNorm code.
MR. BLAIR: Okay.
DR. LEVIN: I'm not giving you a good answer, I guess?
MR. BLAIR: Well, it almost sounded like what you're saying is that you are getting pretty close to providing information that might be sufficient for e-prescribing vendors to just take the Structure Product Label and do the drug-to-drug, drug-to-allergy or is this something that I don't understand? Can they do it? Would they be able to do it just as well off of the information in the structure product label?
DR. LEVIN: Well, in the Structure Product Label would have the ingredient. It would have mapped to the National Drug Code. So both of those are in the Structure Product Labeling. If the information vendor wanted to get that information, they could get it right from those two pieces if that's what they're using to, say, “If you're allergic to this ingredient, don't take these products.” That information would be in the Structure Product Labeling because the ingredient is in the Structure Product Labeling. So you would know every drug that has that ingredient in it.
MR. BLAIR: It sounds like you're saying yes. So maybe I'll do the corollary question. What is it that the RxNorm code provides that would not be available directly from the Structure Product Label?
DR. LEVIN: It provides a single code for this clinical drug. That is not in the Structure Product Labeling at this time. There is no code that says, “For oxycodone and ampicillin” together that are these strengths that are a solution or something like that. There is not a single code in Structure Product Labeling at this time to do that.
MR. BLAIR: Okay, thank you.
MR. REYNOLDS: I had a question on your slide that was potential future initiatives and you made a statement. It was under other drug product data elements, and you talked about NCPDP billing units. And you made the statement that that would be information sent to the prescriber. I would have thought that would have been information sent to the dispenser. So I'll show my lack of knowledge…
DR. LEVIN: No that was my misspeaking. What I was trying to say was this would be a, this communication from the FDA but from the manufacturer we would be passing this information along to the health information system, whoever needs that kind of billing information. Like you said, it would be the dispenser.
MR. REYNOLDS: I'm just trying to keep up with it, and it just struck me as we listened.
DR. LEVIN: And I just thought that that was a good example of something. If you step a little bit further away just from, and this is from the FDA perspective, this is my own opinion, not that that's necessarily FDA's issue, but it's here the Structure Product Labeling is a way that to communicate this kind of between two groups that are using this product and the Structure Product Label could help in that.
MR. REYNOLDS: Oh, good point. It was my edification. Any other questions in the room?
MR. BLAIR: I've been thinking of more as we go along.
MR. REYNOLDS: We'll wait a couple more times until Jeff sits back. He figures it out and then sits back.
MR. BLAIR: I think that we're all aware of the fact that there's an issue that hasn't been resolved with respect to controlled substances being included in prescribing in terms of security requirements for that. So, when I ask this question, I'm not getting at that. I'm not asking about whether they could be used to me prescribing or not, or the security arrangements. But what I'd like to ask is, with the Structure Product Labeling are those also for products that are controlled substances?
DR. LEVIN: Yes, they are. You mean under the DEA control?
MR. BLAIR: Yes.
DR. LEVIN: All marketed drug products is under this idea of the drug listing. And in one of the data elements actually that we capture is the schedule. If they're under a schedule in DEA, we capture that information—which schedule. And that's one of the data elements that's in the Structure Product Labeling so you can see which schedule the drug is under.
MR. BLAIR: Thanks.
MR. REYNOLDS: Okay, anything else? Stan did you have anything else?
DR. HUFF: No, I don't.
MR. REYNOLDS: But I would echo Stan's comments earlier. This has been exciting to watch this happen. You guys have done a lot of work to get this to the point that you're going to issue a regulation. We had a sidebar during the break that it might be helpful since this rule is coming out, and we got our meeting at the end of December, and that kind of ends the comment period that hearing the positives that Stan had to say and all of have seen through this work would be to at least put something together in short letter to the secretary commending this effort and recommending that it continue and be supported and that it appears from the things we've heard over a long period of time that it's a real positive step that's going to make a difference.
DR. LEVIN: I'd also like to thank the committee providing a lot of vision and support for this kind of activity. It's been very helpful for us to move forward, so I want to thank the committee for that help.
MR. REYNOLDS: Okay. Any other comments? Randy, thank you very much. Thanks for being willing to be flexible, to make your presentation early so we could kind of coordinate both of them. Thanks.
Agenda Item: Subcommittee Discussion
Okay what we're going to break into next is some subcommittee discussion. I kind of started off with some of the things that we have coming up, and we've been trying to work this last night, this morning, at break trying to figure out we've got a lot going on and how to make it happen. Let me go through a list of what it appears that our inventory is. Right now, what we need to do and just how we appear to have it assigned.
The CHI letter that we're going to produce is being coordinated by
Margery and Stan. So they will be getting us out a draft on that to have something ready for the full committee in November. We have discussed out of yesterday's testimony by WEDI and so on. We have discussed a letter that we could put together for the November full committee relating to what we've heard so far and alerting the secretary as to the current status, as to the current discussion about transition periods. And that is something that we will monitor in our next hearing. But that is something that probably needs attention right now to make sure that it is on the radar screen and to make sure that it does get the appropriate attention from the department and others as to how it's going to play itself out. Jeff, is that a fair stipulation of what we said?
MR. BLAIR: Yeah.
MR. REYNOLDS: Okay. The third, obviously, is going to be a Hippo Report that will be out to us soon. Sorry, back to the NPI letter. Karen and I will coordinate the preparation of that draft letter for the committee. The Hippo Report will need the involvement of everyone. Because obviously it's based on the hearings and the information that we heard over a substantial period of time and what we've seen happen, so we will want to be editing that… Jim Scanlon, his group. So Jim's got that. It's just that we will then have to group up once we get a draft of this out and approach it accordingly. The NHIN, yes this is not the NHIN committee, but it really constitutes most of the NHIN committee. So Simon, Jeff, and myself are coordinating that effort and letter that's being produced there. So that's one of our collateral that we're dealing with. You just heard me make the comment on the NDC rule, and I hate that Randy just walked out, because what I was hoping was that, I know that FDA can not comment on FDA's letter, but a little bit like Margery, I would like if Randy could help take the lead maybe with Karen, you, or Denise to put together a draft letter for us to send to the secretary on this rule that's out on the NDC codes so that we could at least, again, commend the effort, commend the fact that it looks like from discussions we've heard and what we saw and what we've kind of driven over time.
MR. BLAIR: I think what we're saying is that it's the letter relating to the NPRM for the structure product…
MR. REYNOLDS: That is correct. That is correct. Thank you for that stipulation. Randy, we assigned you something to help us with while you were out which is we would like you, if you would, to work with Karen's office to help us draft. We know it's not coming from you as a member of FDA. It is you as support to our committee helping us put together a letter that we can get to the secretary from the committee. Now once you help with the draft, so that we can keep you in the appropriate position, then we would have you kind of extract yourself from it so that that way you would not have a conflict of interest as we adjudicated the letter. You think that's a good way to handle it? Appropriate use of staff, but we will not keep you in the forefront of it. We'll take it over. Okay? In other words, you would help us as one of our staff members, but we'll take it over and be responsible for it and would not be you as the author, so you wouldn't have to worry about that. So that way, I think that would be helpful. Because you've sat through all the testimony with us through this whole thing end to end, and you are our subject matter expert in that, and so I think that would be very helpful rather than us kind of slosh through something that turns out in the end not really depict in the subject as it should be. And again, we can overwrite that draft it's not what we had hoped for. So if you would be willing to help us with that, that would be good. And we would want to have that ready for the November hearing with the full committee. So again, we got a lot of collateral we got to get done, again, which this committee seems to be our life as far as this is concerned. Michael?
DR. FITZMAURICE: But would it make sense to, in that letter, to ask the secretary to consider putting both the NDC code and the RxNorm code on a package, given that there is a regulation coming up? Have him look into that since RxNorm links to a lot of drug information that might not otherwise be available to a purchaser?
MR. REYNOLDS: What I would recommend is that you're passing that on to Randy and Karen as the drafters to put it in there. Again, we're drafting a letter and then we're all going to look at it and decide how far we do or don't want to go. I mean, everyone of these issues we have, ya know, once we see a draft, we're all going to have opinions. And then we play it in some kind of open forums, there will be people with opinions. I don't have any concern about adding anything to it right now. If that's a reasonable… and then obviously, it will be decided by the committee whether or not…
DR. FITZMAURICE: Yeah, that's what I'm asking. Okay.
MR. REYNOLDS: I think that's a good way to do it, don't you?
MR. BLAIR: Yeah, I think with Karen drafting the letter, and Randy providing advice, that's a good way of handling it.
MR. REYNOLDS: So Michael, we will ask them to take that into consideration if you would send them any wording you would think would be appropriate about that, that would be great.
DR. FITZMAURICE: Also, do we want to complement FDA, NLM, and NCI for their moving forward drug vocabulary and encourage the free exchange of information among those three to improve patient safety through drugs. I didn't sense that there was any problem, and I think Stuart said that NCI stuff is in the Medi Thesaurus, but it recognizes the cooperation and encourages any additional cooperation that might be needed.
MR. REYNOLDS: Any issue from anybody else? Randy, you good with that? No comment? Alright, good. Okay.
Anything else on that particular subject before I move forward?
Okay, we've obviously got to produce a number of things before November. We also have to start planning '07. And we, ya know, kind of really had to be focused on some key initiatives in '06. And so we've gotten away from our kind of our idea of a balanced portfolio where we look into the future. We look into the near future, and then we kind of maintain the current. And so what I would like to recommend is, I would assume that in the November committee meeting, we have a break-out time. That we use that break-out time… I'll send out ahead of time the chart that we've been using on what our portfolio needs to be and the items that we need to think about. Send that out ahead of time and then use our break out time out on the full committee meeting to adjudicate what we would want to make our '07 look like. Does that make sense? Jeff, you good with that? Everybody okay with that? Because I think that's the next, trying to do that now is not going to be productive because we don't have it out.
DR. HUFF: Yeah, that's fine with me.
MR. REYNOLDS: You good with that, Stan? Thanks, okay. Yeah, let's do that so we can really get a good look at '07. And at that point, please bring your calenders because then we'll start trying to set our meeting dates for '07 and we'll be in good shape as far everybody having their dates with them, and we'll make sure we get what we need done.
Okay, let's talk about December. We had a December hearing. We have discussed it some yesterday, some last night, some this morning, and some a few minutes ago. With the fact that we're sending an NPI letter now, and then we would want to closer to the date probably weigh in again. We also would want to be able to hear a little more about the e-prescribing because it would have passed the end of the year, and maybe people can talk to us. That we would have some kind of a January hearing. Marietta, if you would start pulling people as soon as you can to look at a January hearing where we would definitely probably want to hear more of a diverse testimony on NPI then just WEDI, and to make sure the record's clear, WEDI has done an excellent job of grouping up. But one of our strengths is to make sure that even though some people are grouped up, we hear from everyone, and we give a complete and balanced look at that. Because again, we have a February full committee, so then we would be able to put another letter out in February that could really start discussing exactly what May is going to look like and whether or not this extension should have anything wrapped around it that we would want to recommend if there were such a transition period. Comments on that?
MR. BLAIR: If we're going to have a meeting in January together, more information on readiness for the NPI where we would be making a recommendation in February. Does it add a lot of value for us to have a letter in late November?
MR. REYNOLDS: Well, the discussions that have gone on is that since we have not said anything, but we've had a couple sets of testimony. And for example, WEDI's initial letter went to the secretary. We have now heard twice from WEDI, and I think we would be remiss. I mean, the discussions have been, I think Karen agrees, we would be remiss at not making sure that the secretary is aware that this is not just an easy walk to May. So from us, as the advisory. If you look at the results of the surveys that were presented, the industry is quite a ways of from being able to actually fully meet a May date. So, I'm not sure if that makes the answer any better.
MR. BLAIR: Maybe my only question was: If we're going to send a letter on this issue in February, why are we sending another one the last week in November or like December, ya know two months before.
MR. REYNOLDS: Because it's six months from implementation. At that point, it'll be two to three, which means… In other words, right now we're talking about May, and a little bit past that. In February, we're going to be talking about what does it really look like end game is?
MR. BLAIR: Okay.
MR. REYNOLDS: That's the thought. Now again, what is important in each of these? Let's do the draft letters. Remember, this committee has had letters withdrawn. We have withdrawn them. We have done whatever. But it's this idea of if we have a message, write the letter and send it.
MR. BLAIR: It's a heads up to the secretary that there is an issue.
MR. REYNOLDS: Yes, that's what this is. Very short one Jeff.
MR. BLAIR: Okay.
MR. REYNOLDS: Just to say that an issue has come to
our attention. We know it has come to his attention in other letters. Industry we've heard from some groups. We think that he needs to work with CMS and others to monitor this as well as we will and in February we would give further guidance. So, that's kind of what we're thinking.
MR. BLAIR: Okay.
MR. REYNOLDS: But again, let's look at the letter. Let's look at what it says and what it does and whether it does. And if it doesn't, then we obviously withdraw before we even submit it. Some of things are getting to a point now where things are moving so fast that the timing may not be exactly appropriate, but waiting for the next one may not be appropriate either. That's the hard part of some of this journey that we're having to go through, especially with us doing it and then going to the full committee so we kind of get aligned with what that is too. Stan are you good with what we've been saying so far?
DR. HUFF: Yeah, I'm fine.
MR. REYNOLDS: Okay. So that's our plan that we've talked about and kind of come forward with. So, I'd like to hear any other comments from anybody on them. We will be talking. We've got plenty of action going on. It's pretty exciting actually. It's good to have these many subject that are making a difference. It's better than just sitting here and not having anything happen. And there are a number happening right now that may make it look like it's fun.
MR. BLAIR: Harry, I think you've outlined a good play.
MR. REYNOLDS: Okay. Alright, Stan did you have anything else to add to the meeting today?
DR. HUFF: I guess the question I had that Marjorie and I is who is going to initiate the action there?
MR. REYNOLDS: Marjorie is initiating the action and using you as a subject matter expert to make sure that we cover it appropriately since she was on the committee, you would be our filter. One as a subject expert and two as a member of the committee to make sure that the letter is discussed at a committee level. So, since she's on that actual committee, just a little bit like…
DR. HUFF: I wasn't waiting for Marjorie to call.
MR. REYNOLDS: Yes, that's correct. And she said she will have you something within the next two weeks at the latest.
DR. HUFF: Great. I'll wait for that.
MR. REYNOLDS: Okay. Michael?
DR. FITZMAURICE: Could you review the actual items. I apologize for not being here.
MR. REYNOLDS: No, that's fine. Here we go. One, CHI letter- Marjorie and Stan. NPI letter- Karen and Harry. The Hippo Report- All of us. Because that will be done by Jim Scanlon's office and all of us will review that. The NHIN- Simon, Jeff, and I are on that. The NDC rule would be Randy and Karen. That's it. That's what we were saying. So we had a shorter than usual hearing and a longer than usual to-do list. So, maybe we ought to have longer hearings from now on. Might work out a little better.
Anything else anybody has to say? Or anybody from the audience have anything they'd want to add?
DR. CARR: Harry, it's Justine, I'm sorry I wasn't able to be there for some of the hearings yesterday, but if there are handouts or electronic summaries, I'd appreciate getting them.
MR. REYNOLDS: Okay, Marietta will send that to you. I didn't know you were back on, Justine. Are you satisfied with our list of to-dos?
DR. CARR: Yes, I am.
MR. REYNOLDS: Are you upset that you aren't on one of those.
DR. CARR: No, I'm not.
MR. REYNOLDS: Alright. Thank you, and meeting is adjourned. Thanks.
[Whereupon, at 11:50 a.m., the subcommittee meeting was concluded.]