Hubert H. Humphrey Bldg.
200 Independent Avenue, SW
Washington,
D.C.
List of Participants:
Barbara Starfield, M.D., Chair
Jeffrey S. Blair
Simon P. Cohn
Kathryn L. Coltin
Kathleen A. Frawley, J.D.
Lisa I. Iezzoni, M.D.
Vincent Mor, Ph.D.
Clement J. McDonald, M.D.
George H. Van Amburg
William Braithwaite, M.D.
Marjorie S. Greenberg
Robert Moore
TABLE OF CONTENTS
Page
Call to Order 1
Plans for National Provider I.D. 3
Plans for National Payer I.D. 61
P R O C E E D I N G S (9:35 a.m.)
DR. COHN: I want to welcome everyone. This is a meeting this morning of the K2 work group on data standards and security. This is one of the work groups on the Subcommittee on Health Data Needs, Standards and Securities of the National Committee on Vital and Health Statistics.
The focus of the meeting today is on reviewing plans for national provider identifiers and national payer identifiers. I am looking forward to an interesting morning of discussion and hopefully getting the sense of the committee and what our thoughts are on the basis of these discussions.
I want to start with the self introductions by the members of the national committee and staff, and then we'll ask people in the audience to also introduce themselves if they would.
Let me start. My name is Simon Cohn. I am a physician and clinical information system coordinator for Kaiser Permanente. I am also a member of the national committee, and in John Lumpkin's absence I am chairing this session.
DR. BRAITHWAITE: I'm Bill Braithwaite from ASPE. I am on staff to the subcommittee.
DR. DETMER: I'm Don Detmer. I chair the National Committee on Vital and Health Statistics. I'm at the University of Virginia at Charlottesville.
DR. MOR: I'm Vince Mor. I'm a member of the committee. I'm from Brown University.
DR. IEZZONI: Lisa Iezzoni from Beth Israel Deaconess Medical Center in Boston.
DR. MOORE: Bob Moore from HCFA.
DR. MC DONALD: Clem McDonald from Indiana University, a member of the committee.
MS. COLTIN: Kathryn Coltin from Harvard Pilgrim Health Care, Boston, member of the committee.
MS. GREENBERG: I'm Martha Greenberg from the National Center for Health Statistics, and acting executive secretary of the committee.
DR. SCANLON: I'm Jim Scanlon from the Office of the Assistant Secretary for Planning and Evaluation here in HHS. I'm the executive staff director of the committee.
MR. BLAIR: I'm Jeff Blair with IBM Healthcare Solutions.
DR. KRAMER: I'm Michael Kramer with the National Center for Health Statistics.
DR. ROTHWELL: John Rothwell, NCHS.
DR. PEPPER: Michael Pepper, Assistant Secretary for Planning and Evaluation, HHS.
DR. LANDON: Bruce Landon, representing Blue Cross/Blue Shield Association.
(Whereupon, introductions from the audience were performed.)
DR. COHN: Thank you. Obviously, the main focus of the day will be on -- or this morning, will be on discussion of the national provider I.D. We have asked Karen Trudel to make a presentation on the current status and plans for the national provider I.D. After that, I'm sure we'll have some questions for you.
DR. TRUDEL: Good morning. Actually, I wasn't planning to make very much of a presentation, because I know you have a lot of questions. We have sent out some materials a few weeks ago. They included answers to some questions that were raised at a previous session, I believe in February, and a revised version of the preamble for the provider identifier regulation, which we hope adds some information and possibly clarifies some other issues.
I was planning on just taking your questions. Is that all right?
DR. COHN: All right. Are there questions from the committee?
DR. MC DONALD: I read your rather extensive material. There are a number of things that are left in that as A or B -- we will do this or this -- and I didn't understand when that will be resolved or how that will unfold.
DR. TRUDEL: Yes, there were a number of issues that we felt that we didn't have sufficient information on to actually say we would like to do it this way. Rather than put out one possible position and not appear to be open to all of the options, we determined that it would be better to simply raise some options, invite comments. After the 60-day comment period, we will review those comments that come back, and then in the final regulation that is published after that, there will be decisions on those points.
DR. COHN: I guess I'll ask one of the A or B or C or D questions. As I was looking at the location code issue, or the handling of locators, was an issue brought up by the National Committee in January. I found that the variety of choices you had were even a little confusing to me, to be quite honest. I'm wondering if you can review those, as well as the pros and cons of each of those options.
DR. TRUDEL: Okay. This is probably the most perplexing and controversial issue that has to do with the provider identifier. I and some other members of our team attended a WEDI(?) workshop last month, and I know Rich was there, discussing issues about the provider identifier. It became obviously apparent that there are a number of different points of view, even within the industry, as to what a location code should be, and when and how it should be used.
One of the conclusions that the group came to at the end of two days of really exhaustive discussions was that this was an issue that needed to be discussed in the context of the claim standard, because the use of the location code in claims processing was -- and the myriad needs that are associated with that were what was really driving a lot of the controversy.
Some of the issues included the fact that payers occasionally today use provider identifiers and location codes for a number of different purposes, even within the Medicare program. In some cases, a carrier will assign multiple provider numbers, not only to determine which physician is practicing at which physical location, but which tax I.D. the service is being provided under, so that the payment can be credited to the right 1099. In some cases, it indicates which physician is practicing at which location, and who is going to receive the payment. In some cases, the location code or the entire provider identifier is used to designate a physician practicing as a part of a group, at a group's location, and perhaps he then has another identifier for himself practicing at that location on his own and not as a part of the group.
This is why it is so confusing. We have had a lot of trouble reducing this to a few simple choices. The proposals in the regulation are essentially to not have location codes at all. We would gather information about practice locations, we would gather the addresses. The legislation clearly tells us that we need to take into account multiple practice settings.
We could assign a two-position location code or each physical practice site that doesn't already have an NPI for another purpose. For instance, if you are practicing in a hospital, the hospital has an NPI. Do you need to give the physician a location code to attach to his NPI to show that he is practicing at that hospital? Some people think so.
Some payers feel that if a physician is practicing as a member of a group, he should have a location code that could be attached to his identifier for each location of the group where he practices. We found out a little bit about, for instance, how the Mayo Clinic works. When you start taking into account all the various practice settings and the thousands and thousands of physicians, it becomes unwieldy. However, some payers feel that that is a necessary. They really need that information.
So those are some of the options that we have been hearing.
DR. MC DONALD: The location code I think is -- what you have described, though, I think is some ad hoc solutions to problems poorly analyzed.
DR. TRUDEL: I agree with you a hundred percent, Dr. McDonald.
DR. MC DONALD: What you really have is, you have got a person, and that is not ambiguous. That is pure. There is no confusion. It is well-defined in the documents that it is a human person. Then you've got the groups and other things which are tougher, but I thought also well-defined.
I think if you're going to name a thing semantically location, that is what it should mean, shouldn't it? That makes it easiest and less confusing, rather than have it be the group sometimes and this sometimes.
I know the mechanical problems you will have in trying to do real locations independent, because addresses change, and you can have one side of the building that has a different address than the other side of the building. But it would be ideal if one could push the insurance companies to get to the analytic problem they have, rather than their solution to the problem, to see if you just couldn't get a pure location and your NPIs, and then everything would be kind of clean.
DR. TRUDEL: I agree. The identifier group met last week with some representatives from the implementation team that is working on claims standard. We talked at great length about what provider numbers are used for and what kinds of information you really need to process a claim. In other words, do you need the whole address in order to determine what price and localities are, or could you really do that on a nine-position zip code? What exactly do you mean when you say I need address data in order to do fraud and abuse checks, or medical review checks? Do you need a full address? Do you need a zip code?
We talked about geo codes. Nothing is off the table at this point. What we intend to do in partnership with the claims implementation team is to prepare a white paper on provider issues that relate to the claim. We hope that we will be able to post that on the Web site that you are using, and obtain some comments that way.
In the meantime, we feel that there is enough flexibility in some of the options that we are proposing in the provider I.D. regulation that it would be acceptable to continue on, publish the notice of proposed rulemaking, and begin the comment period, even without having absolute final solutions to some of these issues.
DR. BRAITHWAITE: Clem, I was at that WEDI meeting as well, and I found it very educational. When they finally got to the point where they all understood that nobody understood what the other person was saying, they started drawing a primitive entity relationship model and came up with four different entities, two different kinds of locations, and eight relationships between them that they were trying to represent with two fields, that is, the NPI and the location code.
Those included the provider as a person, the group that the provider who actually provided the association was associated with at the particular time that that service was provided, the facility in which the provider was providing the service, the person to whom the money should go when the service is actually paid for, and then the physical location of where the service was being provided and the physical location to where they should send the money.
All of those things were being confused and trying to be pulled into this very simple explanation, because that is the way we used to do it, and that is how much space there is on the form. In fact, it is how some of the people seem to be thinking about it.
So I agree; this needs a great deal of analysis and thought by the industry, about how to apply this thing.
The only positive aspect that came out of that was that all of the payers who were there at least felt that given what HCFA has proposed to implement, they could figure out how to use that to do what they do now, even though they would all use it differently.
DR. SCANLON: Karen, I wonder if it would be fruitful for the NPI itself, for the national provider identifier itself, if you would just describe, in addition to the associated data items and what makes up the pure national provider I.D. per se, the associated items, and then -- what I gather is, that location code is not necessarily proposed to be included in that set of items that make up the NPI, but potentially some other item or some other standard.
Could you just run through, from the number itself, its nature, to the associated data items that would make up the core of the provider I.D.?
DR. TRUDEL: Certainly. There has been some confusion also about whether the location code was part of the NPI itself. We have actually moved all discussions of the location code to the part of the preamble that talks about the supporting data set.
The NPI itself is proposed as an eight-position alpha numeric identifier, of which the last position is an isostandard check digit. That is the standard that we are proposing.
We are further proposing to -- because some payers especially have indicated that they have problems right now dealing with all numeric, possibly all alpha numeric identifiers, we are going to modify the algorithm that assigns the identifiers to run through most of the numerics first, and then to begin assigning the alphas in a controlled way, to give folks additional time to work with whatever software problems they may have.
But the NPI itself is in a position alpha numeric which incudes a check digit. The data that goes along with that is partially demographic, date of birth, place of birth. We have a mailing address, where we can communicate back with that provider if we need to obtain additional information. The ever-popular practice location address information with possibly an associated two position alpha numeric check digit, which would be -- if for nothing else, for simplicity of processing that data within our system so that we don't have to address matches and checks every time we want to use one of those elements of data.
We have education data for certain types of providers, physicians, for instance. That is school and year of graduation. There is board certification information, state license information where that is applicable to the provider type. We have the ANSI XWEL proposed taxonomy that is being worked on now by a subgroup, and that is essentially a very exhaustive description of all types of providers and contains three levels of detail, and other numbers that that provider has been known by, if we know them, the Medicare UPIN, the social security number, which would be voluntary for an organization, the employer identification number, if the individual has one as a business.
That is essentially it.
DR. MC DONALD: You mentioned the double check digit> What was that? It went by too fast. You said you have a two character check digit>
DR. TRUDEL: No, no, a one position check digit.
DR. MC DONALD: That is on the provider I.D., but when you're talking about the data. You said the data was the address location?
DR. TRUDEL: Yes, and associated with that address might be a code that we would assign to that address, just a two position code.
DR. MC DONALD: That is the same two position --
DR. TRUDEL: Right, that is the location code that is so popular.
DR. MC DONALD: But I thought you said something about a two alpha character check digit, and that went by very fast.
DR. TRUDEL: I'm sorry, perhaps I misstated that. No. If I did, that's not what I meant. Location code. That is the location code.
DR. MC DONALD: I was hoping that it was an actual address code.
DR. TRUDEL: No.
DR. COHN: Just to clarify, were you referring to the elements in the national provider file just now, or was that something different than what you were describing?
DR. TRUDEL: No, that is what will be in the file.
DR. COHN: In my review of the documentation, there was also some information about sanctions? Am I mistaken?
DR. TRUDEL: Right. There is a code to reflect the type of a DHHS-OIG sanction and an effector date for that sanction, if there happens to be one, and an end date if the sanction is in the past. Those data elements at this point are being discussed as whether they will remain in the database or not, which is why I didn't mention them.
DR. MC DONALD: It is worth discussing. I think for the sake of standardization, to tangle up the standardization issue with the political issues is a tactical mistake. If you just standardize it and then you worry about your policy and enforcement issues independently, that would be wiser.
DR. MOR: There was a lot of language in here about the ability to then link across state sanctioned histories, because not everything that is a sanctionable event yields some kind of HHS-OIG action, but rather as a state action. I didn't hear anything about history. Many people are licensed in multiple states. How for instance would that be indicated?
DR. TRUDEL: We are essentially assuming that we would be tracking a provider throughout its existence, and therefore, if a provider moves from one state to another or even for instance begins their career as an RN and then becomes an MD, they would not receive a new NPI, and all of that back history will be associated with their record.
DR. MOR: So this is not just archived, but continuous update?
DR. TRUDEL: Well, --
DR. MOR: How? Is it in multiple records?
DR. TRUDEL: It is. It is. Essentially, some of these items have start and end dates associated with them, the practice locations. For instance, you could say this location is no longer active, and this location now begins to be active as of this date. So all of the data will be maintained in the database.
DR. MOR: So if somebody is practicing in Washington, D.C., Virginia and Maryland all at the same time, they have relicenses and three indicators all there?
DR. TRUDEL: Yes, and all active, right.
DR. BRAITHWAITE: Karen, it might be useful for people, if you went through the list of items again and talked about those items which were there for the purpose of uniquely identifying a provider, versus those items which were there to help the system work.
As an example, the licenses. I'm not sure that whether you are licensed in the state of California or Maryland is useful to uniquely identify a person or provider, but I may be wrong about that. Maybe it would help for you to go through that list again.
DR. DETMER: And if it is not too complicated, maybe it would be. But if you could also reflect what you are seeing out in the general system, in addition to HCFA's issues as they relate to those things.
DR. TRUDEL: Perhaps I should begin explaining a little bit about what our search algorithm does. It is essentially a two-part process that looks at certain data elements and makes an immediate assumption that you cannot have, for instance, two different people with the same social security number.
We also make an assumption that two different people cannot have the same state license number in the same state. So that is one of our absolute duplicate checks.
If a record passes all of those duplicate checks, we then begin to look for duplicates in a scoring algorithm using a soundexing of the last name, the first name, part of the date of birth. I believe part of the place of birth is in there, and if gender is available, that is also in there.
At the end of this mathematical weighting process, a score is produced. If the score is below a certain amount, we assume that there is no possible duplicate on the file. If the score is over a certain amount, we assume that there is a duplicate on the file, and we will tell whoever is intending to enumerate this person, wait a minute, this provider is already here, and the score is over 80, or whatever. You must look at this record and determine that it is not a duplicate. Then anything in between is a possible match.
We again return all of these possibles to the enumerator in order of likelihood, and ask them to sift through this data, determine whether one of these is indeed a match, and then to take the appropriate enumeration action. So those data elements that I have mentioned are all used in the search algorithm for individuals.
For organizations, we are also using the address and the employer identification number. I am frankly not as up on the organizational part of that search algorithm as I should be.
Some of the other data elements that we have included are partially for the use of the enumerator in tie-breaking. If I have three possible providers and I know where mine went to school, I can look at that data on the other providers to determine whether perhaps it is the same one, so that some of the information, including the practice location for individuals, are very important for enumerators in the tie-breaking process.
There are also data elements that are very useful for payers and others who are attempting to match NPIs with their existing provider files. If we didn't maintain that practice location, the address data, it would be very difficult, for instance, for a Blue Cross plan to say that this Dr. Jones is the same as their Dr. Jones, when they know where he practices. So some of that information is very useful for matching purposes.
The UPIN and the prior provider numbers are a prime example of what others will need to have in the file in order to be able to make it usable.
Yes?
DR. MC DONALD: Along the same line, do you have fields to represent when a record, a provider, NPI number is no longer valid, and then what the valid number is? Because you will get duplicates on them.
DR. TRUDEL: Yes.
DR. MC DONALD: I hope you're not assuming you won't get duplicates on them.
DR. TRUDEL: Oh, no, we know. We know we are going to have problems, and we know that we are going to have to be able to delete duplicate numbers.
DR. MC DONALD: No, you don't want to delete them.
DR. TRUDEL: But we want to sure -- or de-activate duplicate numbers. We want to make sure when we do that that we link the de-activated number to the active number.
DR. MC DONALD: So you have two fields then. I didn't hear them mentioned. One to say this is de-activated, and one to say what the correct number is.
DR. TRUDEL: Right. There are a number of data elements that are internal to the system that we haven't mentioned here. There are start and stop dates for a number of these items. There are certain flag dates and fields that we use for internal processing. We haven't mentioned any of those in the regulations or in any of our paperwork reduction items, because they are internally generated.
We also maintain a date of death. I think I forgot to mention that, for individuals.
DR. DETMER: Let me ask it a different way, perhaps. What percent if you will of all of this issue do you anticipate to essentially seek a lot of correspondence between what HCFA does and the general industry does? Which pieces of this -- you mentioned the location codes as one issue, but what pieces do you anticipate are likely to be the most worrisome as this rolls out?
DR. TRUDEL: The data that is required in order to enumerate an individual, we don't anticipate is over and above what most payers require in order to pay a provider. It is fairly minimal information. You have to know who it is, you have to know where it is, you have to know the fact that he is authorized to provide the services that you're going to pay him for.
So generally speaking, we have not heard either providers or payers indicating that the amount of data that was required was onerous. Some of the data elements we know are going to be not available from some payers. We know that, because some of the data elements aren't available for initially loading the database, even for Medicare. The open registry does not have, I believe it is phone numbers. They don't collect information on gender.
So we are trying to plan for a database down the road, where we will have the information that we need and we will probably obtain it on a flow basis.
DR. DETMER: Part of my concern on this is that obviously, the goal of this legislation was to simplify. Obviously, our recommendations, if they end up generating huge costs, don't quite meet the objectives that we are all about, either for the government or the general --
DR. TRUDEL: Right. Aside from the sanction data, I don't believe I have heard any comments from the industry as a whole that certain data we are gathering, we should not gather. In fact, at the WEDI working session, there was discussion that there were additional data elements that we would have in the NPS database in the field itself that we had not anticipated making available to the general public, that they felt would be helpful.
DR. MOORE: How representative at that WEDI session was the industry there? Did you hear from a representative sample of everyone? How well are we doing the job for the industry? We are doing it well for Medicare, Medicaid and some of the federal programs, but how well is the provider database going to represent what the industry needs in the commercial side to do their business?
DR. TRUDEL: I would say we have gotten a very representative sampling of payers. And in terms of providers, there were not a lot of provider organizations represented at that particular session, but we have made a real effort, I think to discuss these issues with various industry professional groups.
About a year ago, when we believed that the NPS would be implemented for Medicare only, we did actually go through the official Paperwork Reduction Act notice in the Federal Register. We did receive a number of comments from AMA, ASIM, I believe, I'm not sure which organizations. But we did go through the 60-day comment period, and did not receive a lot of significant comments about the data burden.
DR. MOORE: Just one follow-up on that. I received this week, I think yesterday, the WEDI writeup that came out of those two day sessions. I'll get that to the staff, and it can be shared with the members, because they are going to issue a letter, isn't that right, Rich, to the Secretary? Identifying what they believe will be the additional requirements that they are going to ask for in the legislation.
This was all trying to be done prior to the reg getting out, so that some things could be put in the reg, and knowing that we won't have time, as we are moving along very rapidly on this.
So there are things happening from the industry as we unveil what it is we are trying to do. I think there was a very open exchange at the two day session, with about 85 people from the industry participating the payer and provider I.D. If there are any specific questions about that, maybe Rich can handle them, since he was one of the people there from the industry side.
MS. COLTIN: In the proposed rule, there are a number of options presented for what entity would actually do the enumeration. I didn't notice in any of the for and against arguments that were presented issues around the actual validity of the data elements that you are going to be collecting, things like medical school, year of graduation, board certification and so forth.
Do you have any plans in that area for validating the data, or will that be a factor in determining what entities do the enumeration?
DR. TRUDEL: Yes, actually it is a very important issue. One of the reasons that we want to keep the number of entities that do the enumeration fairly limited is for exactly that reason, to assure data quality and control over the database.
I believe there is some mention -- and perhaps now that you have raised this point, I need to go back and see if it needs to be discussed in greater detail. One of the jobs of an enumerator is to assure the validity of the data. We did not go into a lot of detail about, does that mean that you check the license with the licensing agency, or you request a true copy of the license, or exactly how you do that, what the evidentiary requirements should be.
But yes, that is very definitely one of the things that we expect enumerators to do.
DR. MC DONALD: The NPI and those associated variables needed to manage the file, that is one thing, as you said. These other attributes are another thing. I wonder how you think they will be used. That is, addresses and telephone numbers and those things change, especially as the phone companies change the area codes.
So are you thinking in terms of, everything would go in this master file, and that is the only place insurance companies will get the data? What I can envision, the insurance companies are going to go separately for that data, anyway, and there will be lags. So that maybe shouldn't as much a focus, at least one ought to think about where the focus is for this national part. And HCFA might have its own requirements.
DR. TRUDEL: Right. There really are twofold requirements. The first obviously is to assure the validity of whatever data we are using as part of our search algorithm, because the heart of any enumeration system is its ability to determine duplication. So obviously, the social security number if we get it, we are going straight to Social Security Administration to check that.
DR. MC DONALD: No, that is not what I was asking. I think you've got two classes of variables to deal with in the database, one class needed to identify the individual and keep your database clean. The other is needed for reasons that might be useful, like address and telephone number, other ones that you enumerated, I see those things as not saving the other payers' time, because they are going to go directly to the providers to get those things as they need them.
So I would like to know how --
DR. TRUDEL: I agree, but there is a caveat there. Where I was going with my statement was that obviously, the most important evidentiary checking is for the data elements that we are using for duplication.
However, if we are gathering data that others are using to for instance associate the proper NPI with the proper provider in their file, we do feel that there is a certain need to keep that data as current and as valid as we possibly can.
I agree with you that that is probably a lesser requirement, and that they probably will go straight to the provider. But if we have data that is absolutely inaccurate, and there is an inability to associate the NPI to use the data because of that data not being valid, I think we have a --
DR. COHN: Let me ask a question, just to help clarify. My understanding is that Medicare and HCFA would likely use this database for paying bills, right? Am I mistaken about that? Or are they all going to be using the provider file as reference information on where the remittances go?
DR. TRUDEL: The provider data that they are receiving through the national provider system will be a subset of the data that they use to maintain their provider file and pay their claims. For Medicare purposes, the data that is being gathered on the HCFA 855, the provider enrollment form, a subset of it will be enumeration data and the rest of it is Medicare enrollment data.
DR. MOORE: There will be a lot of other data that is required in the claims processing system by their contractors that is far more extensive than we would be carrying in this. Like, rates for different periods and all that, those are not part of this.
DR. TRUDEL: Right. Medicare however does intend to use the national provider file as the official source of that subset of data elements. So at least for Medicare providers, any changes that are made to a Medicare provider will come through the national provider system before they go to a Medicare provider file.
DR. MOORE: What I was trying to clarify was that, if the database is being used for something, then it likely will be very good quality data, like the remittance of claims. If it is not being used for anything, and just sort of sitting there, it is likely not to be very good data, and the phone numbers and addresses will likely go out of date very quickly.
DR. TRUDEL: Right. The database itself will not be used directly by Medicare in the payment process, but they are treating it as the trusted source of the data.
DR. MC DONALD: Maybe I might not have been clear. Is the Medicare payment side separate? That is not all providers. The NPI is all providers. Is the model in your mind that insurance companies and other payers will come to you for data? My bet is that they are still going to go the wrong way, and so that is why I think we ought to be careful about the added fields, just because someone might want them, because they are still going to go out and get them.
The second thing is, when you mentioned that the big effort is to make sure that they can use their own internal provider numbers, as they have been in this mapping process, it also raises some flags. Does that mean you envision that insurance companies will still require payers to use the insurance company's provider number and then they will translate them and send them to you, or something like that? That is, these other numbers should go away, I would hope.
DR. TRUDEL: Yes, they should. I was speaking about the period of time during which NPIs are going to have to be appended to existing provider files that don't have them. What one would assume at some point in time when this is fully implemented is that providers will have NPIs, and that issue would fall away.
DR. MOR: This discussion leads me to wonder about Medicare agencies, some of which are going to have more difficulty than others, making these kinds of administrative adjustments, and whether this provider file, if thought of as being updated and a resource then, might do part of that job for the Medicaid agency, in terms of being available for billing claims, et cetera, as a service.
Which I think gets to Clem's question. If you intend to have this as a service function, it has got to be up to date. Now, we can understand all of the rationale, but if you assume that every state Medicaid agency will do this on their own, then yours is going to be as a last resort will not be updated.
Do you want to comment on that?
DR. TRUDEL: I would expect that this is a valuable service, and I would expect that it would be used as such.
DR. MOR: So you would imagine that the state of Maryland would actually maybe even read off of the site on an ongoing basis updates?
DR. TRUDEL: Yes.
DR. MOR: So that it wouldn't make much sense for Blue Cross/Blue Shield of Maryland not to do that. I guess that is the point.
DR. FRAWLEY: I was going to say, it raises the whole thing: what if they don't? What kinds of reinforcement mechanism is there?
DR. MOR: It is really the amount of work. This is much enlarged from NPI. NPI is great, and I am a hundred percent behind it. There should be no controversy, or little, associated with that.
But this idea of this file, someone has got to get clear in their mind about how it is really going to be used. I think that really, you're going to have more layers in your process than the typical direct -- if you want the bill paid, send me your latest address, you don't get it.
So what I worry a little bit about is, on the one hand, if this was the live file for everyone, that would be great, except then you may have some privacy issues. On the other hand, I worry that everybody is still going to get their own, anyway, and so it may be more work, for these extra fields.
DR. COHN: Further comments?
DR. MOR: That is really the point. What is the position? It is going to cost a lot of money to keep this up to date and active. If it is a service and thought of in that way, that is one thing. If it is just there if you want to use it, that is a whole different mind set.
DR. MOORE: But I don't think the legislation requires anyone to use the database behind that. The legislation requires that they use the unique identifier that is being proposed.
If I were -- and I would daresay that in the development of the Medicare transaction system that we are going through in HCFA on the Medicare claims, we are going to create our own provider file that that system is going to use, that will take data from here and supplement it with other information that we need to process that claim. No way would I just download everything, bring it over and create my own file for claims processing, because I'm going to make sure what information I have, how I use it, is maintained in an accurate fashion.
I don't know that I would -- maybe Karen could kill me for this, but if I am running the system out there, I am not going to rely without some checking that I don't have the right address to pay that provider. In the case that there is not a change out there on the add list, the address I have in here for that check going out to someone is going to be as right as I can, because it will not pay Clem McDonald if it goes to Don Detmer. He is going to be mighty unhappy.
DR. COHN: I was just going to comment though that what we are describing I think is some of the vision of rthe payer I.D., how they saw using the database. I think there is some question here about whether or not this is a very good database and keep it updated, about whether it needs to be out there.
If we have this national provider system, but you have your own database, and you are updating it, but not updating this file, this file is going to become very poor quality very quickly. Those are some of the questions.
DR. TRUDEL: If I can raise one other issue, we are not putting a requirement on the enumerators per se in the regulation to maintain the data. But we are proposing to put a requirement on the providers to report the changes, and to then have them keep that data up to date in the system, so that the enumerators essentially would be a conduit for that data, and the requirement is on the provider.
DR. MC DONALD: I think there are two hopes for this database. I think we've got to at least get them really clearly separated. One of them is, they will tell who the providers are, plain and simple. That is the big goal.
The second one, get all this other stuff that might be useful. I think as you push for the second goal, you get more into this angle of, can you keep it up.
Because you have to have these other steps in the process and these checks in the process, my guess is you'll get a slow person with that. The yellow pages will know about the address changes before you will.
In that context, I think you might at least want to put a circle around those variables that you really want to enumerate, and then those variables that are like these other extended ones. If people aren't going to use those as primary data, you maybe oughtn't to make as much strength on that as you are claiming to keep those up.
DR. DETMER: It seems like what I am hearing is, at least we know we need -- to mandate the identifier, sufficient data to know that in fact we have actually identified them. So that is a core set of expectation you have to meet.
I guess in the spirit of the legislation, beyond that, if there are other things that do then create some implication, it seems to me if we can really feel that they generate to really do that, we ought to consider it. But the first task is probably the primary one. Then I think these others are important, but probably secondary.
DR. TRUDEL: Right. Are there some data elements that we have discussed that you feel are going to be particularly difficult to --
DR. MC DONALD: Telephone number and address.
DR. DETMER: I would assume also that in addition to mailing address for checks, the most important thing is going to be bank accounts to which funds are going to be deposited.
DR. TRUDEL: Right, and we have none of that. That is obviously something that is program specific, and may in fact differ from Medicare to Medicaid or whatever. So we would not be gathering that information.
So it is the address data and similar -- phone number and that sort of information that you feel could be problematic?
DR. MOR: Also, the licensing I.D. If someone gets a license someplace else, I don't know how your system is going to have that updated.
DR. FRAWLEY: Karen, there is another point, too. If the provider retires or is deceased, how would that notification -- they are good points.
DR. MC DONALD: You have to tell HCFA before you die.
DR. FRAWLEY: That you are planning on dying. For some people it may be when we get this implemented.
DR. COHN: Other comments on the data elements?
DR. MOR: I'm not sure whether this is the locator code or the different I.D.s, I'm not sure I necessarily see the difference between a nursing home owned by a chain. At any given point in time it might be owned by one company having a separate I.D. number than five different practice sites owned by another group. But they seem to be treated differently in terms of the I.D. numbers at the organizational level. Did I read it correctly?
DR. TRUDEL: Yes, you did.
DR. MOR: Why?
DR. TRUDEL: When we attempted to define a provider at the beginning of this entire process, we thought, this is going to be the easiest thing. Then we will just go on to the hard stuff, like figuring out what the identifier is going to look like.
The more we began to do it -- and this was a group of individuals from various government agencies and Medicaid state agencies. I know Marjorie sat in on some of those discussions. It became very obvious that we could identify a living person. That was pretty easy. We knew what he was, who he was and when he went out of existence.
When we began to try to define organizations, we discovered that it was almost impossible to do that without referring to programmatic rules. Like, Medicare says a hospital is this, Medicaid says a nursing home is that. There were layers of organizations that some plans recognized and some didn't.
The compromise position that we came to, that we found would be workable for the most number of people, was to enumerate an organization at the smallest, the most granular setting, to allow the various plans, to build those together in whatever ways they needed to identify them.
We then had a definition of an organization that we thought could work. We had a definition of an individual that we knew would work. The question was, what about groups of individuals, clinics, group practices, a variety of business agreements and arrangements that are sometimes driven by billing constraints? That is why we established groups, because a group was a combination of individuals and didn't essentially have its own existence, aside from those individuals, in other words, a partnership of physicians isn't actually authorized to provide health care, the physicians who are licensed in that state are authorized to provide health care. We treated groups the same way we treated individuals.
So the real distinction is between the individuals and the organizations, and then the groups follow along with the individuals.
DR. MOR: So the groups don't exist?
DR. TRUDEL: Did that explain anything?
DR. MOR: So groups really don't exist?
DR. TRUDEL: Yes, that's correct.
DR. MOR: The reason I bring this up is system this system for provider I.D.s is supposed to exist and make sense for a long time to come, and there is one thing that is a guarantee, that provider rearrangements and arrangements, even within a physical setting, like a nursing home, are going to be changing dramatically, even more so than they are right now within the same physical space of a nursing home, physical therapists, occupational therapists, laboratory -- roving mammography units and actually even hospice sub-units are providing distinct and billing distinct under part A and part B. God only knows what they are billing for in the private sector.
So the utility of the organizational umbrella, particularly given the fact that they get bought and sold fairly frequently, you might want to think about that, or at least imagine it. I don't have a solution. It is just --
DR. COHN: I think we have about 15, 20 minutes left before we take a break. I actually wanted to direct the work group, rather than to continue to ask questions to Karen, I think it is important that we begin to get a sense of where we are in terms of this provider identifier and see what sort of consensus we can convey to the National Committee as a whole about a recommendation in this.
MS. GREENBERG: May I just ask one quick question? I agree with you, that is very important to do, so I don't want to take up time from that. But I was looking at the -- you have referred to the taxonomy a few times. I was looking at the minutes, I think it was from the January meeting, whoever presented it, I think it was Joe, that you provided a copy of the taxonomy to the subcommittee. I wasn't sure that that had happened.
I think if you could do that, I think that would be helpful.
DR. TRUDEL: Yes, definitely. If you don't have one -- in fact, it has changed since then, and I believe will be changed again very shortly. I believe there is another ANSE meeting coming up at the end of this month.
What I am told by our representative is that this time, this is it.
MS. GREENBERG: I know. I was involved a while back, but I think it has changed. I think that could be reviewed at the same time that the NPRM is going forward, because it is not really a part of that.
DR. TRUDEL: Right. I understand that the changes that are being made are not significant. They are not changing the structure, or any of the major elements. I think some of it is just rewording. If you would like, I can get you a copy of what we have now.
MS. GREENBERG: Yes, because even with some of these elements that were being mentioned, although I agree that the main goal is obviously to enumerate each person uniquely, and that this would be a major step forward in standardization, there is some hope that the system might be able to be used as a sampling frame for various types of other provider studies that now require various and sundry ways of identifying providers.
So I think that that might be a reason to keep some elements in the system, even if one recognizes that their accuracy is -- or validating them is of lesser importance to the system. But I think that if we strip it too much, it will lose is utility in that regard, and I do think we should keep that in mind.
DR. COHN: So I guess as an action item, we will get a copy of the taxonomy?
DR. TRUDEL: Yes.
DR. COHN: Okay. Now back to my question.
MR. BLAIR: A question. You were saying that you were waiting for some kind of approval at ANSE. I didn't understand what that was.
DR. TRUDEL: That was on the taxonomy. I believe they are close to having a final version of that.
MR. BLAIR: When you say ANSE, --
DR. TRUDEL: This is the X12 working group.
DR. COHN: Yet I try again here. Are there any other questions before I try to move the focus of the discussion slightly? Okay.
It seems to me that there are a couple of areas that we may want to have a consensus about. One of them is the NPI itself. The other is the associated file. I think those are really the two ways that I am conceptualizing it. Is everyone in agreement with that sort of conceptualization?
Having said that, what is the sense of the work group on the NPI itself? Are there issues?
DR. MC DONALD: I think you can't talk about the NPI as a separate field. You've got to at least take on that part of the associated file that gives you enough identification.
DR. COHN: Let's try to combine these things pretty concisely, rather than trying to -- because I think that is a second discussion.
DR. MOR: And I think it is actually possible to disentangle those. It is an historical process. One is, you need as much information as you can have in a structured set in order to come up with an unduplicated count and enumeration. After that, it is a totally separate issue as to whether those data necessary to come up with, or attempt to come up with an unduplicated count should have an investment of being updated. So one can differentiate those.
DR. FRAWLEY: Yes, the update issue is where I think most of us are struggling with this, how does that process come together. Otherwise, all of this good work then can fall apart.
DR. MC DONALD: That is not true, though. If you are just enumerating, the only update you have to worry about is duplicating and adding new people. It is not the same as trying to keep up all the fields and addresses and telephone numbers.
DR. FRAWLEY: Right, that's what I was --
DR. MOR: That's the point.
DR. FRAWLEY: That is what I'm concerned about, or the person that decides to retire, or the person that is deceased. This guy has got a license in another state now, or in another practice. There are just so many variations on that theme, of how do you keep these files updated and current.
DR. MOR: I think the NPI per se makes all the sense in the world. I don't understand the location code, I haven't understood it from the very beginning.
DR. MC DONALD: That is not part of it.
DR. MOR: That is not part of it, then that is great.
DR. MC DONALD: Everybody is nodding, so I guess everybody is for the NPI without a location code at least. I'm not saying it can't have a location code.
DR. MOR: It should be separate. It is a separate data element.
DR. COHN: Okay, is that the sense of the work group? Okay. Now let's talk about the file. We have already heard it being split a couple of different ways, that which is required for the enumeration and identification uniquely of that individual versus additional data elements. Discussion?
DR. MC DONALD: I think they should make it as small as they can, for the sake of simplification.
DR. COHN: Well, I would at least like to have a discussion. I don't know if there is any data on the relative accuracy of, let's say, HCFA files or government files on providers or insurers or something like that, and the quality of data that the TRW and the like maintain. Otherwise, this information is essentially a commercial issue. There will be a provider, a vendor that will extract and update all this information. It is really just a question of what that vendor will be charging or making available to other payers, billers, including Medicaid agencies. So it is really whether this is a governmental service function for the world of payers or it is a commercial function.
DR. MOORE: What are you saying? Are you saying, how should this be maintained?
DR. MOR: Yes. Should it be maintained in the private sector or should the government maintain it?
DR. MOORE: I just wanted to clearly understand where you --
DR. MC DONALD: There are two variables in this. One is how wide the database is, how many fields you have in it, and the other one, who works it. I think we should make it as small as is feasible, in terms of wide number of fields.
But it depends a lot on whether the industry would plan on drinking out of it -- my bet is, they won't, because it won't be able to be as direct, and they have these other issues they need to get from the providers anyway. If that is the case, then I think we should definitely keep it tight. If that is wrong, then make it broad. But maybe the audience knows better what the commission will want to do.
You've already said that even HCFA is going directly to get data from the providers.
DR. MOORE: There are some things that we will need from the provider that will not be there: where will the check be deposited, and electronic funds transfer, rates that we established based upon our coverage determinations and our rules for paying, all those kind of things. So that is a separate issue, which will not be part of this.
DR. MOR: But Bob, let me ask you. The rates are all basically table driven, so some characteristic of that individual, that provider, and the event that that provider generates in the form of a claim, are used to select the data element of a piece on the table.
Now, everybody out there in the whole world is going to have different tables for prices. But are they going to base them on different bits of information?
DR. MOORE: Yes, possibly. That is what we found when we met with the WEDI group. They have multiple variables that they have used a provider number to uniquely identify what they were going to pay. That was the sole purpose for having all those different variations of the provider number. That is one of the problems we were having, is having the provider identifier service too many requirements simultaneously.
DR. MOR: I guess my concern then is, in terms of administrative simplification for any given provider submitting information, there will still be potentially quite different batches of data that you will have to submit to one insurer versus another insurer versus another insurer, based on the array of data that they want to have to draw a particular price out of a table, separate altogether from CPT codes or what is actually done.
DR. COHN: Other comments?
DR. DETMER: I guess I want to come back -- since this is the hearing where we are talking about this, obviously, if we know enough about the provider, you don't necessarily need an alpha numeric number also.
Just as a matter of discussion, I see the value of an alpha numeric number, but I think the question is, to me, I think we are going to be likely wanting the number plus the set of data close together as the core. But the question is, do we agree or not that that is what is needed and if so, why, and if not, why. Do you follow my question?
DR. COHN: Yes. You are asking a very fundamental question which goes back to -- we all nodded our heads about an NPI, and you are raising the issue of, they really don't need an NPI. Is that what you're asking?
DR. DETMER: Yes. You are using another data set to confirm that the number you assigned it is in fact that provider. If you are setting those criteria to decide that that number you assigned is in fact that same person, do you need that number.
DR. COHN: Discussion by the group?
DR. MOR: You don't want to have to go through that process, counting in the scoring and so on and so forth, every time, having done that, trying to match --
(Simultaneous discussion.)
DR. DETMER: That is my sense of the answer, but I was interested in having it on the table.
DR. COHN: I think it is the sense of the group that we do need an NPI. At least, that is what I was hearing from everybody. I think the discussion we are dancing around right now really has to do with the issue of the national provider file as a minimum required for unique enumeration versus adding other things to do. That is the discussion that we have been having.
I am hearing -- now, we have not specified what is required for unique enumeration, but I am hearing a sense that, yes, we think we need that. Is that a -- we'll start with that one?
DR. FRAWLEY: Yes.
DR. MOORE: I think if you think back to Marjorie's questions about the availability of a file to those more than just payers for information needs in the country, where would one go to say how many doctors of a certain type do we have? Do you run around to all the payers in the country and count them up and add them up and total it? Do you find out what kind of health care is delivered in a certain location, how many doctors live in what county or practice in what county in Virginia versus another place in the country? How many hospitals are located, what size?
Where do you get your information about what health care is available in the country? Do you go and survey every payer? If you have some information, as Marjorie said, it may not be a hundred percent accurate for payment, where you want to know that you are paying to a physician or to a place correctly, in the right bank and all that. But if you were doing information, and you start using that information and doing your analysis of the health care delivery system, which is not the purpose of this file perhaps, but it is a utility of the file, as Karen was saying, then you start looking at what is the cost of having that information there, how many times do people use it, how many times do they go out and ask the public for it, or other places for it, et cetera, et cetera.
DR. MOR: That is why I wanted to ak the question about, is this the government or a commercial function, the updating process. The government function should I think from this committee's vantage point, for the whole committee, at least have some mechanism for providing a sampling base.
I can tell you the difficulty of sampling from providers in the current world, but the federal government is -- we are doing -- AHCPR and most of the other federal agencies are putting a lot of money into a big study, a national study of HIV. The sampling frame that was necessary to do that was just horrendous, because of the problem of not even having a single physician, because nobody is a member; all the members of AMA and all those kind of fun things.
It is impossible without this very, very difficult process to go across state licensure lists. They're all computerized and so on.
So the information necessary to actually come up with a unique enumeration is I think very important and valuable from a research perspective. The real issue then is where is the responsibility for updating it. Should it be an every three year updating responsibility that is an activity funded out of the National Center for Health Statistics, as strictly a research function? Or does it have value that other people would drink from if it were updated.
I think that is the fundamental question.
DR. FRAWLEY: Right, and I'm not sure that we have had enough input on that point from industry. We have talked to different groups or heard from different groups on that point, but --
MS. COLTIN: I think that there is at least one other function that would make it valuable for private organizations to drink from it. If the data were accurate, if they were well validated and up to date, to help support the credentialing process.tialling process.
When you think about administrative simplification from the provider's standpoint, they have to be credentialled by ten different health plans in three different hospitals and eight different nursing homes and whatever. The forms that they have to fill out, the validation that has to be gone through again and again by different organizations, if in fact these data were validated and could serve as a credible source for that sort of function.
I think that private organizations would probably not only have an impetus to supply good quality data because they would be wanting to use it back, but might even be willing to help defray the costs.
DR. DETMER: I don't think it is inconsistent with the charge of the committee to talk about the recommendation relative to the identifier, and then what kind of process is appropriate to make sure that it in fact is usable and relevant, valid, at least from my perspective. I don't see that that is in the scope of what we could do.
DR. COHN: Yes, I think certainly that is the sort of discussion we are having. Knowing from my own experience of the amount of information required for credentialling, --
MS. COLTIN: Even if it is helped.
DR. COHN: Yes, I was going to say, this is a small piece of it, a very small piece of what credentialling is all about. I would certainly be myself concerned --
DR. DETMER: But I guess what I'm saying is, if we are talking about a unique number here, it is only unique if it is, and you know it. So to me, I think that is a part of what is in the concept, is the number. Otherwise, I don't think it does simplify anything. So to me, it seems like if we have a recommendation for a number, having some process to try to know that that is indeed the number, is not inconsistent at all with this whole process.
DR. MOR: The whole notion of how this is maintained or sustained -- I don't know enough about how different the credentialling process is from area to area. I know that I have to write all these silly letters for providers who have worked with me from time to time, and I'm not even a provider.
But that is one thing that you could at least do. For instance, an entitlement check is now done automatically from every place in the country for a Medicare beneficiary. You could imagine an entitlement or eligibility check as the first kind of thing that a potential provider could do, and then begin the process, then do the more detailed process afterwards.
We have entitlement files maintained and sustained by the government, whether it is Social Security Administration or Medicare. Is this a function for government, or is it a function for somebody else? I think if it is, then there are certain elements of information that you would want to maintain on an ongoing basis and have updated by having it updated in the same kind of way entitlement data are for individuals; are you entitled to be a provider.
DR. COHN: Does that include sanctioned data?
DR. MOR: One can imagine it slightly differently, but yes. I guess I would make a comment on that. I think it is important that this be a good database and be utilized. I actually think, at least to my view, one of the defects of the notice has to do with the fact that there isn't a vision of how it would be used within the aspects of commerce.
But having said that, I am also very, very concerned about any sort of sanctioned data of any sort without strong issues of privacy, strong issues of authentication, all of the things that will suddenly turn this thing into a 300 or 400 page notice rather than a 50 page notice.
DR. DETMER: I think there is reason to walk before you run, too.
DR. MOR: That is well said.
DR. DETMER: The issue is, how much is a walk, because we do need to move it. Not just a step. I think we need to get something that actually moves.
I am uncomfortable thinking that we can really go jog right out of the chute.
DR. FRAWLEY: Yes, that we're going to be doing baby steps first.
DR. DETMER: We're just crawling right now.
DR. MOR: The issue for instance of validating board certification is also not a simple process. It would be wonderful to have a single location where board certification data are actually all located. But we don't -- providers have no difficulty about having their version of board certification made public and available, and yet, that is probably as subject to error and distortion as is information about sanctions.
DR. SCANLON: Before we leave the point, the general point was that the file presumably, if it were high quality and reliable data and could be maintained as such, would have as one of its uses a research and statistical use. I think I am hearing you say yes. I think the subcommittee is having that as one of its major uses as well.
DR. COHN: Yes. I guess I am hearing a number of things. I am hearing, number one, that everyone is in favor of having at least in the provider file enough to enumerate that specific individual. I don't think anyone is in disagreement with that, is that correct?
DR. FRAWLEY: Correct.
DR. COHN: Because we obviously need to be able to uniquely identify that person.
DR. MOORE: And you all feel comfortable with the approach that we are talking to do that? I would like to make sure that we --
DR. MOR: It seems to me that the only other alternative wold be to have first the provider -- the provider I.D. would be -- the consumer I.D. and the provider I.D. would be the same. That is the only other option. It is, as you articulated, an eight-digit number that would not be an individual health insurance number.
DR. DETMER: I guess what I would say relative to what has been reported, both from -- not just the government side, but also the discussion generally, it seems like this is not an unreasonable approach, on the basis of what we have heard today, at least.
DR. COHN: Now, let me ask the question again, because we are talking about two sides of different things. The question that Bob had was, is the national provider file an appropriate -- we have already talked about the NPI. I think we can put that discussion to rest. But the question is, is the national provider file in the right direction, fine the way it is? Does it need a little more work on it, from my comment about sanctions, which I think you have also commented on, Karen.
I would just ask the committee, what is the sense of the committee regarding sanctions? Am I the only one who has some concern about that?
DR. FRAWLEY: I concur with your remarks.
DR. COHN: Okay. Now, are there other elements?
DR. BRAITHWAITE: Excuse me, Simon. For Jeff's use, people should say something instead of just nodding, and also for the transcript.
DR. COHN: Thank you. Please remind me on that one. Now, are there other elements that we think should be off the plate? Off the plate, should not be part of the national provider file?
DR. MOR: I don't think you want to have billing information.
DR. TRUDEL: It is not there. In broad strokes, I think what you're talking about is data elements that relate to education, data elements that relate to board certification, --
DR. MOR: Licensure.
DR. TRUDEL: Licensure we have in the search.
DR. MOR: So the results would continue to be in the file and updated?
DR. TRUDEL: Right, and the practice location address, phone, e-mail that go with it. I think that is pretty much it.
DR. MOR: If the address is not updatable, I suppose. From a research perspective, if you have an address once, we can begin tracking this person down and looking for it. E-mail will probably change often until we each have our own phone number that we will walk around with, which is probably not too long from now. But if you have the fields, it make sense over the next ten years, the communication will evolve to that one component.
DR. SCANLON: Simon, on the sanction information, the principle here being not to include that among the core -- the number itself or the core data items, the principle being that it is a unique programmatic application versus an item needed for identification and enumeration generally?
DR. COHN: Are you asking me that question or are you making a statement?
DR. SCANLON: The group. It is a question. Why did we conclude -- what is the principle on which one concludes that the sanction information is not part of the -- we have to deal with this internally, obviously, but --
DR. COHN: I can make some comments. Kathleen, it looks like you wanted to start.
DR. FRAWLEY: You go first.
DR. COHN: Okay. My comment would be that it would be a unique program application that you more readily sit within the Medicare transactions or something like that. It also has major issues of both sensitivity, issues around authentication, issues around privacy and confidentiality, that far supersede any of the other data elements we're talking about. Those are the reasons that I would speak against having it within the data file.
DR. FRAWLEY: I concur with you. Obviously, we would want to know if somebody was dropped from the Medicare program or the Medicaid program. I just don't know how you set that in a box, so that it is available, that a physician has been dropped from participating in programs.
DR. BRAITHWAITE: Where would you want that?
DR. FRAWLEY: I don't know. I'm trying to figure out, how do you build a box where maybe -- do you have to go through another layer?
DR. BRAITHWAITE: That is already public information. Our Inspector General publishes that information every year. So it is public information in that sense; it has already been released.
DR. MOR: If you don't want to include sanctions, it would probably make sense not to, for all kinds of reasons. Then as long as you have this unique provider number, then all other databases including sanctioned information in their various locations, would have to be committed to translate to that for linkage, doing whatever.
Right now, sanction information is publicly available on all Medicare-Medicaid certified nursing homes and home health agencies. An awful lot easier than it is to get information about that about various other kinds of providers. You don't get in on hospitals, but there are other databases available for that kind of information.
DR. DETMER: My concern was more logistics. I think the time might come where it might be totally appropriate. There are enough kinds of different sanctions and why this sanction and not that sanction and how you should track it, as well as trying to keep these others updated. So my primary concern is logistical.
I can see in fact arguments to be made for accountability purposes for it to be there, but I think that is how much you could put in on this at the start, is my concern.
DR. COHN: Is that a fair sense of where we go from here? I would like to give the committee a break, so I am trying to wrap this particular part of the discussion up. Are there any final comments?
DR. DETMER: I would just hope that it would be nice actually to see a listing really, so that when it goes to the full committee, we really do have a precise kind of list of what we have included in there. I'm sure you're wanting that, too.
DR. COHN: You mean, within the provider file?
DR. DETMER: Yes, in that file box.
MR. GREENBERG: I'm George Greenberg. I work with NASBE. I am just curious if you have any recommendations on how this should be maintained, because there are several options in the draft rule, and who should maintain it, and how should it be financed. It seems to me those are big issues.
DR. MOORE: That was a question that came up at the WEDI. Rich happened to be the one to make the presentation. There was a considerable discussion about how fees could maintain it, but no one came forward to offer the capital investment to build it, so people could begin using it, which is a considerable expense, to put all this together into a resource that one has the access.
If I recall what the industry wanted, they want 24 hours a day, seven days a week access, is that right?
DR. LANDON: Payer I.D.
DR. MOORE: Was it payer I.D.
DR. LANDON: That was the one I reported on.
DR. MOORE: Okay. Knowing how much access there would be, this drives one of the requirements for that system that is going to be operational.
DR. COHN: Comments? Let's make sure we've got the last one handled before we move on to this one, because it is a very valid topic. But are we okay and complete now with the file? Are we okay with that?
DR. FRAWLEY: Yes.
DR. COHN: Do people want to spend just a couple of minutes identifying which option that they think is the best option for maintaining it, as well as how to pay?
DR. DETMER: I would like to just weigh in. I think there are three issues here. All these issues, the identifier, the issue of classification and nomenclature, medical terminology, all these things, need to define the language and get some way of defining it. A server that can get that to the users on a reliable basis, and a process by which you continue to look at that and refine it, update it and link it to the past over time.
Now, those are essential functions that I think we're talking about on almost any dimension of these things. So it seems to me that we really as a committee should speak out to that. I personally would favor a public-private strategy on that. But for all of our talking about that, I must say, I'm not sure that this country does that very well. In fact, I don't think it does do that very well. In fact, I'm not sure it does it.
So the point is, it needs to be done. The functions need to be done. There are dimensions clearly of groups out there that can play pieces of this, but it seems to me that one of the things this committee could do would be at least to lay out for broader discussion what we would think that could look like.
There is no question that, indeed, if there is to be simplification, but even work accomplished, it seems to me like we need to do this.
DR. COHN: I guess I would weigh in myself and comment that I personally believe that accommodation of federal programs in a registry would be the way to maintain this piece. But I actually would agree -- actually, I personally think that the strategy as enunciated for payer I.D., which is making Web access, making this database available and actually used in the process of -- potentially a fee being charged for usage might be a useful strategy for which the government could maintain the costs of the database, as well as that would help insure the quality of the data, since it would be used in everyday work. It would make much less likely that people would move or change addresses and not have it be updated, or zip codes or telephone numbers changed without these things being updated in a reasonable fashion.
That is my personal opinion about how best to see it happen. I agree with Don, that I don't think public/private things work very well in these circumstances.
DR. DETMER: Pieces of it may work that way. I just think that this whole set of things though needs to get laid out, because right now it is not a part of our working machinery.
DR. COHN: Yes, and there can certainly be some value added pieces to this whole thing. Other comments?
DR. MOR: Yes. The provider I.D. is different than the payer I.D. The provider I.D. presumably is a much bigger nut altogether. There are many, many, many more providers. There will be more updating of that process.
I still think -- I believe the principle is probably the same. The question is whether there would be sufficient demand for some kind of pay per drink approach for using the information by the world out there. It is unlikely to be used in the same kind of way for commercial business applications because people will just simply download and have their own provider information for any different payer group, and states would have it.
So it is probably more, something should be retained, but it doesn't have to be there for ongoing daily commercial access, but it could be -- some kind of charge could be levied for annual or quarterly updates of it, if everyone agrees that there is an independent agent presumably with no ax to grind, that would maintain and update the information, and then make it available a la the census. The Census Department makes a lot of data available to the whole world, and people use that, and it is very inexpensive to update.
DR. COHN: Other thoughts, comments?
MR. BLAIR: In terms of the definition of providers, it seems as if the countries going through a trend where more alternative health care providers are being considered. Is the identifier in the field restricted to those that are currently reimbursable by Medicare-Medicaid, or is there the flexibility to accommodate a broader definition of providers over time?
DR. TRUDEL: There is definitely the flexibility. The listing of providers, the taxonomy that we have been talking to, goes well beyond the boundaries of Medicare and Medicaid now, and there is no reason why, assuming that any given provider can be accommodated in the context of an individual, an organization or a group, that alternative practitioners couldn't be inenumerated.
DR. COHN: Thanks, Karen. We have somehow gone to a final question here, which is obviously the enumerators as well as financing mechanisms. Is there a sense of recommendation from this group around some of those things, or is that an issue that we would want to bring up with the full committee for further discussion?
DR. TRUDEL: Could I make a request, Dr. Cohn?
DR. COHN: Yes, what?
DR. TRUDEL: If there is any language that we can add to the preamble, either to explain some of these issues more clearly or elicit specific comments from the industry at large? I know you mentioned that we don't know enough about some of these things yet. I would be happy to do that, if you could just let me know what those would be.
DR. DETMER: When do you anticipate this next round of information being available? You said it was the third iteration of some responses, if I heard you.
DR. TRUDEL: That is the taxonomy. The regulation itself, you have the most up-to-date version of the preamble, and we are currently making sure that the regulation text matches it, and working on the impact analysis. When those three things are done, we will be ready to move forward again, hopefully very soon.
DR. COHN: With that, I'm going to suggest we take a ten-minute break. We're running a little late, I believe, significantly late, and we'll come back together at 11:15. Karen, thank you very much.
DR. TRUDEL: Thank you.
(Brief recess.)
DR. COHN: Why don't we get started? I want to apologize to everyone for the relatively abbreviated break, but as you know, we are obviously running a little late.
I want to thank you both, Mr. Pagels and Rossiter, for coming and joining us to talk about the payer I.D. Why don't I just let you both start?
MR. PAGELS: Thank you. We are pleased to be here today. I am particularly pleased that Faye was able to make it. Today is not only tax day, but her birthday.
We are the other identifier, but we basically are on a parallel track to the NPI, that is, the payer I.D. unique identifier for health plans, and employers that offer health plans. That is how we have been presenting the payer I.D. project.
I think over the past several months, the biggest appreciation we have gained for the payer I.D. project is that it is not just a Medicare initiative, but it is an industry initiative. We continue to consult with industry. Most recently, we talked to WEDI in March at their policy forum and gathered some good information there.
One of the things that we are learning is to keep it simple, keep it as high level as possible, make it usable. I think as we go forward, that is something to keep in the back of our minds.
Basically, you have been given the information similar to the NPI. We are basically here to answer any questions that you may have regarding the materials that you received.
I guess one other basic concept we need to keep in mind is that really, the payer I.D. initiative as it initially was conceived, and even today, is to replace the alpha numeric -- the way in which address information is currently maintained in different formats, in different files, using different names and addresses, or addresses that aren't always the most up to date, and to have a file that is accessible to the user community, to the commercial community, and one that is going to be useful to them, and up to date.
We are ready to take any questions you may have. I hope we're ready.
DR. COHN: Questions from the committee? I guess I'll start out with one basic question. As I review what was -- what you had provided us in the way of information, I was having some confusion in my own mind, only because I was thinking that a payer I.D. number and file might be very useful in the whole process of coordination benefits. Yet, I couldn't quite understand how this would really be helpful in all of that, other than being able to tell someone where to call. Can you describe that for me, how this might assist with that overall process?
MR. PAGELS: Well, I think the basic premise is, the idea itself is not a processing system. What it is, is an identifier that for one thing, we get insurance companies and health plans to come forward and get enumerated. So we have now a registry of information that is centralized, and it is hopefully up to date and useful information.
So when folks are coordinating benefits across insurance companies, they can basically identify them. They can go to one source for that information, and they have the most up to date information in terms of address, information in some other characteristics of EDI mailing address for that entity.
So I think the benefit of payer I.D. really is that it simplifies the maintenance of that information for the industry, and makes it more usable. In that way, they are able to more quickly identify the other insurance company.
The real benefit of payer I.D. will be its use on claims forms, its use on different health care transactions when the industry at large starts using it on their EDI transactions and sharing information, getting it to providers, so that when providers -- and getting it perhaps on health insurance cards, so that when a beneficiary or a patient goes into a physician, that physician can quickly identify the location of that insurance company, the name of the insurance company and where to send that bill. Then if there are any other coordination of benefits that has to go on between insurance companies, that information is again available in a database, and hopefully up to date.
DR. BRAITHWAITE: One of the questions that came to me was some confusion about payer I.D., because people are trying to differentiate between plans, a specific contract for health care, and a plan on an organizational perspective, like a payer. So you've got Blue Cross, that has 1700 different plans, and then there is this payer I.D. Sometimes like for an organization like Blue Cross, you are actually giving a number to Blue Cross, and then you are giving them an ability to add a bunch of numbers to the end of it. Are the numbers that you're adding, or letting Blue Cross add to the end of this plan I.D., if you will, related to the specific contracts? Perhaps you can go over how that fits in the real world, so that people are not so confused.
MS. DIETRICK: We are still deciding on the idea that the suffix is there. It is one of the things that we were going to propose in our MPRM and to the committee members here on how far should we be involved in letting the companies enumerate. Should we enumerate high, and let the claim come in the front door and let that company do its own little thing? Or should they be allowed the opportunity to have the numbers for different plans or different process locations, should they all be the same, should they just be only processing locations.
We need some answers on that, because we definitely don't know which way to go.
DR. BRAITHWAITE: That explains the confusion.
MS. DIETRICK: We are trying to look at this as -- as you said, it is a base plus two digits, the suffix being the indicator for a processing location, a type of policy. We are trying to make it fit the best we can with the needs that are out there.
We have also added to the data elements that we sent to everybody. What we are going to add is an actual data element that indicates the type of plan it is, whether to an HMO or fee for service. That is not on a current application, but will be.
MR. PAGELS: I think this came up at the WEDI meeting. The first day during the WEDI presentation, one of the companies came forward and said, we think we need not just a hundred suffixes, but a thousand suffixes. Then by the second day, everybody was saying, well, wait a minute, we just want one number, and our company will act as a gateway, and then we will get it to our front door and distribute it to the different business divisions that we have.
So as we speak, as Faye said, we don't really have a firm approach on that.
DR. SCANLON: Would there be any value, or is it even possible to include not just the type of plan, but as other data items, some of the products of the company, various insurance products, that the company might offer? So not just the name of the credential and not just the overall name of that particular plan, let's say it is a medical policy, but actually some other products? Or is that beyond the purpose or structure?
MS. DIETRICK: That might be beyond, but I'm kind of in the dark here on this one. That is something that we can propose for comments, to see how it fits with the industry.
MS. COLTIN: Have you heard from or talked to any states that actually require this kind of information? I know in Massachusetts, the hospitals that are required to submit discharge abstract data to the state have to code a payer type and a unique payer code. The payer code is really at the broad product level.
I know in the plan that I work for, we probably have 10 or 12 different codes in our payer codes, just in our state, in Massachusetts, and that includes Harvard Community Health Plan commercial, Harvard Community Health Plan point of service, Harvard Community Health Plan Medicare, Medicaid, and several others under a single organizational umbrella.
So this would be less specific as proposed than what our state is already requiring hospitals to submit to the state, in terms of details on payer.
DR. MOR: Jim, in terms of the research issues, I think precisely for the reasons Kathy just enumerated, at least two or three of those product classes within HPHC are from a research standpoint at any given time identical or different, depending on whether it was 1987 or 1994.
So they are highly idiosyncratic, and don't have much intrinsic research value. This is not my area of research, but I know people who are struggling with taxonomies of plans are finding it extraordinarily difficult to come up with anything that is reasonable and decides to stand still for more than six months at a time.
So I don't have any advice for how to do that coherently.
MS. DIETRICK: I think at one time, somebody also mentioned that the same product can be offered by the same company in different states and still be different. So that is a lot of information in one little Band-aid.
MS. COLTIN: Just a clarification. The product is different than the benefit package. I know at least in Massachusetts, there are really only four product types per plan that we code, I'm sorry, five: HMO, point of service, PPO, Medicare and Medicaid.
A lot of that is really -- that is the level that most of the researchers want. They want to know, is this Medicaid enrollee in a managed care plan. They want to be able to differentiate those from the commercial enrollee. They want to know if they should look for claims under a different payer code for this person, because with the point of service option, they can go outside the plan, and there might be a third party administrator for outside claims or things like that with a different payer code, but the person is still under this plan.
So that is the level that they collect it in, in our state at least.
DR. MOR: That is a very useful taxonomy, except for the fact that with a unique individual number, -- you wouldn't really know, but at any given point in time, that individual might be Medicaid covered or might not be Medicaid covered.
MS. COLTIN: The code is on every claim, so you can tell at the time this service was delivered, this is what the coverage was.
DR. COHN: Any further questions on that point? I was actually just going to change the subject just slightly, to just ask if you would talk a little bit about your plans for making this self-sustaining, or how will you see it being used out there. We talked about this in relationship to the national provider I.D. and files, and it looks like you have had further thoughts about how you might be able to turn this into a marketplace success, I guess is how I might describe that.
MR. PAGELS: By marketplace success, I don't think you mean profit.
DR. COHN: I guess self sustaining sounds pretty successful.
MR. PAGELS: Self sustaining, right. I think we see the registry as being that. Right now, it is basically an initiative that we are putting forward in terms of developing the database with NTIS, National Technical Information Service, correct?
At some point, when folks come forward to get enumerated, they will have the opportunity to become registry users, commercial users if you own the data. Those commercial users will have the opportunity to establish an account with the registry, and there will be fees associated with accessing data.
We see two different access methods. One would be an online inquiry, basically. Another method would be a fairly simple method of going in and seeing the information on the file. The other way would be actually being able to download the information and maybe putting it on your PC and manipulating it.
For internal use, you would see -- and external use, those that chose to download the information and use it externally or market it to others, we would want to assess a little bit higher fee. We think that would be appropriate.
So we are playing with the fee structure and thinking about some options there, but the bottom line being that the registry itself be self sustained through user access fees.
MS. DIETRICK: Also, there is going to be either electronic directories or paper directories. There will be a fee for those also.
MR. PAGELS: And those are actually a subset of the total fees.
DR. MOR: A question. It seems to me the value from a user's point of view of this kind of registry would diminish the more aggregated and invisible the unit data are, so that if all Blue Crosses are given a flat digit number and then it is only up to Blue Cross to hand out 20,000 additional numbers, and that is not visible to the potential user of the information at the national level, then I'm not sure what value it has. That may have less value.
MR. PAGELS: Even though we are talking about allowing some opportunity to get suffixes, which is what you're talking about, enumerating different business divisions, if you will, those actually become numbers that are housed in the database.
DR. MOR: Okay. That was not clear.
MR. PAGELS: Yes, so those would be part of the database.
DR. SCANLON: About how many payers are envisioned that exist now at any rate, and presumably would make up the --
MS. DIETRICK: If we enumerate all the ERISA plans, which the last time I talked to Labor was like four and a half million, and all the payers that we are aware of, we are talking about 19 million. But the database has a capacity of 100 million to enumerate. So we have quite a ways to go with the numbers.
DR. SCANLON: In terms of implementing the registry itself, you would start more or less, it sounds like, with the HCFA payers you are aware of.
MS. DIETRICK: Currently right now, we are in the process of loading or populating on database, and we are going with anybody that does business with Medicare right now. It is the easiest ones.
Our primary insurers, Medicare contractors, Medicaid state agencies, we are currently enumerating those. However, we don't have a lot of information on file. We are limited to name and address. Because we have not gone forward with the final rule, we cannot go out there and collect the additional information that entities will need in order to process claims.
So we are hoping that once this becomes final, we will solicit some information and get it then to populate the database. But yes, we are going to start with the current enumeration of people or entities that do business with Medicare.
MS. GREENBERG: Could you explain the extent to which this database is going to enumerate employers? I have some confusion on that. Maybe Bill can provide some clarity on what the requirement is in HCFA to uniquely enumerate employers, and which employers.
DR. COHN: Who wants to tackle that?
DR. BRAITHWAITE: I would be happy to talk about the intent. Clearly, employers have a tax I.D. number which the Social Security Administration assigns to them, and that is a very valuable way of identifying them. But there is a problem both with the privacy concept and with the role that an employer plays at any one particular time. They are not only the employer, but in many cases they are also the plan and in some cases the provider of care at the same time.
The intent was to assign employers numbers that indicated their role, so that the privacy regulations and other approaches could differentiate, this is an employer as plan, that is paying for health care, this is an employer as a provider of health care or this is an employer as an employer, that is, a payer of salary, to try and differentiate those, not to replace the EIN when it is appropriate as the employer per se, but only to uniquely identify the employer when they have different roles that really need to be differentiated from their role as an employer.
MS. GREENBERG: The employer could have a provider I.D. and a payer I.D. and an EIN. This system will enumerate employers who are functioning as payers?
DR. BRAITHWAITE: That was the intent, yes.
MS. GREENBERG: When it says those that offer health benefits, that could imply a much broader role than what you mean. It doesn't mean that they offer to their employees that they can sign up for Blue Cross. It doesn't mean that they are offering health benefits to the employees as part of their benefit package. It is referring to employers who actually are serving as payers.
DR. BRAITHWAITE: Yes.
MS. DIETRICK: We are trying to -- and obviously, we have to work on that a little bit -- we didn't want to replace the EIN, I don't think, at least not with the payer I.D. We don't even have the capacity for that one. But we were trying to get close to the law as it was.
MS. GREENBERG: I think in what you provided us, the word employer was only there once. So maybe that needs a little more explanation.
MS. DIETRICK: The information we did give you was a draft. We are constantly trying to read through -- and any sections that you believe we should, as you just said, we will go back and concentrate on those areas.
MR. PAGELS: Can I ask you for a clarification? When you said only employers that serve as payers, were you referring to like self funded employers?
MS. GREENBERG: That is what I assumed.
MR. PAGELS: Okay, I just wanted to be clear.
MS. GREENBERG: Is that correct?
MS. DIETRICK: Yes.
MR. PAGELS: Yes, to the best of our knowledge. It is an area that we're working in. That is why I wanted to get clarification. I think it is a good point.
MS. GREENBERG: At one point I thought it was much broader, but I understand now.
DR. COHN: A related but somewhat separate question is the issue of casualty and liability insurers.
MS. DIETRICK: In the regulation, I think it is under definition of group health plans, we have added a small sentence in there that says property, casualty and liability insurance. All those will be enumerated. They are included now. The liability insurance people will be enumerated.
DR. COHN: Other questions, comments?
DR. MOR: Did you mention that you're still not sure what the categories of data would be for category of health plan or address set for types of services. Is that still up in the air?
MS. DIETRICK: The categories of health plans? That basically is not up in the air. I think we have pooled all the ones as the law states. The address sets --
DR. MOR: The categories of health plans, that's set?
MS. DIETRICK: I don't think anything is really, really set, but as far as we know right now, I think we have tried to identify as it is in HIPA, yes, and go with that same line of thought.
DR. MOR: I don't know what HIPA is.
MS. DIETRICK: Oh, I'm sorry, that's --
(Simultaneous discussion.)
MS. GREENBERG: Are we talking now about the element that you said you added?
MS. DIETRICK: No, that's a new element. That I think will probably address the HMOs, the fee for service. We already have a spot on the application that handles the Medicare and Medicaid, but we can further specify that we see it as a nice element for research. You can at least zone in on something.
MS. GREENBERG: Are you coming up with standardized definitions? I know this has been a problem. This was a problem over the years in uniform data sets of trying to capture what then was referred to as expected source of payment. The last version of the committee working on this, they felt there was so little -- it was such a moving target, and there was so little agreement on terminology, that it kind of punted.
MS. DIETRICK: I don't see how it would work otherwise.
MS. GREENBERG: I think it would be helpful.
DR. MOR: Have you got those definitions yet?
MS. DIETRICK: No, not yet. We're working on that. They will be in the regs, so feel free to give us your comments on it. But we will be talking to industry in general.
MS. GREENBERG: I would strongly support doing that, as long as we can get some consensus.
DR. SCANLON: And there may be an industry like the AAHP or something. There may be an industry taxonomy that is being worked on, in terms of definitions.
DR. BRAITHWAITE: Keeping in mind that the purpose under K2 for this registry is to simplify the administration of health care, and we presume by that that this would be used primarily for the simplification and support of electronic transfer of information, would you go through the elements that are in the database that supports the payer I.D. and tell us which ones are absolutely required for that purpose and which ones are good things to have for other purposes?
MS. DIETRICK: The primary name of the applicant, that is required.
MR. PAGELS: The effective date would be a requirement, just so they would know at what point they should be sending it.
MS. DIETRICK: And the termination date.
MR. PAGELS: And the termination date.
MS. DIETRICK: On number five, we actually had the applications and we are asking them whether they are going to apply for a single number or block of a hundred. We need to know that, because if they do a hundred, they just can't take numbers and run with them. They have to give us some data back, entered into the database, so that we know what they are talking about. So they are all going to be requirements there.
The description or purpose of the payer I.D., that gets into -- and we may even change this one to add to it, but we are hoping that that would be like, if it was a Blue Cross/Blue Shield plan, they might say that it was a point of service type of -- or an HMO there. So we might have to expand upon that one, instead of adding another plan, maybe expand on this one. But that would be a requirement.
All the information regarding the address, the zip code, all that is definitely a requirement that we need in the database. Entities, if they need to call up and get further information, they need to know the number.
The e-mail, the EDI, they are going to be applicable. The cross references, they are applicable, but if there is any in there, like the name or the alias that would better help an entity determine exactly which payer I.D. to use, if we get some comments that they believe that they should be required, then -- you know.
DR. BRAITHWAITE: I don't mean to belabor the point. I just want to make sure that we have a focus on the primary purposes, and like we spoke about with the NPI, to make sure that it is implemented in the most cost effective, useful way in the beginning, and that other useful things that might be added on and can be justified should be done as add-ons and not try to push it through and load it down with a bunch of -- yes.
DR. MOR: I have a question about -- if an entity ceases to exist or is merged or what have you in this world, some entity with ten different plans, four or five of them may get bought and they are left with three or four of them, or whatever. They might go out in different parts.
Now, they would have different I.D. numbers. They would change numbers and use an NDC code because the prefix would change, or how is that going to work?
MS. DIETRICK: We see it right now, if the number is assigned to the plan entity, and it is the main company and that company is sold out, that number is retired and never used again. However, anything under that, like the suffix assigned, that would follow where that other one went. So it doesn't disappear.
DR. MOR: But it could be a duplicate?
MS. DIETRICK: No, it shouldn't be a duplicate.
DR. MOR: Because if you change the prefix because the prefix goes out of business, and this next guy has ten sub-plans as well, and this one is number ten and this one is number ten and they get joined together, they might have the same suffix.
MS. DIETRICK: But you are going to keep the same original six numbers, the original assigned to -- the base number will stay with that. The payer I.D. in its entirety is nine digits. If the first six identifies Blue Cross/Blue Shield, and if Blue Cross/Blue Shield somehow gets bought out, whatever those first six numbers will follow the product line.
DR. MOR: I don't understand that.
DR. COHN: Why? This seems very complex. I guess I had read it in a statement that you had, trying to take all the intelligence out of the coding system. It sounds to me like you're trying to put it back in.
MS. DIETRICK: No, we're not trying to put it back in. It is just that if a number has been assigned to an entity, and I guess we really haven't thought as far as the individual plotting.
(Simultaneous discussion.)
MS. DIETRICK: Maybe I should retract that, because we need to think about the products being split. We were thinking about companies being merged, and that way the number is retired.
DR. MOR: But they are parcelled out and bought and old and put into different packages now.
MS. DIETRICK: We'll have to think about that one, but I think that is probably true; we would treat it as a entity in the new --
MR. PAGELS: Being a new entity. We have retired the old number and assigned a new number.
DR. MOR: Will it be archived for history, so you can find out --
MS. DIETRICK: There will be an audit trail there, but it just will not be used to enumerate a new entity.
DR. COHN: Just as with the last one, we are trying to ask some questions, but also trying to come up with some sense of recommendations. Are we at a point now where we can move to that second item, or are there other questions that need to be asked?
MR. GREENBERG: I had a really brief question.
DR. COHN: The mike.
MR. GREENBERG: Are third party administrators not enumerated? Let's say I'm an ERISA plan using a third party administrator. The ERISA plan gets the NID, but are third party administrators part of the system?
MS. DIETRICK: We have included them to be enumerated.
MR. GREENBERG: So they will have their own numbers, even though they don't actually bear any insurance risk?
MR. PAGELS: What happens in that situation is, the ERISA plan would come forward and the third party administrator as their agent would have a number assigned to it for the employer. That is probably not answering your question, though. But they would be enumerated.
MS. GREENBERG: Each plan gets enumerated and then each line of business. For the third party administrator, it sounds like each ERISA plan is a line of business (words lost). If you are assigning the number to third party administrator, they are like the Blue Cross plan that has all of these sub-plans or something. Is that right?
DR. COHN: The private conversation is interesting, but I can barely hear it myself. So if there is an inclusion here, you might like to restate what comment there was that would be useful for the taping and otherwise.
MR. GREENBERG: It sounded from that description that you were treating the third party administrators analogously to what you just said you were doing for Blue Cross plans. Basically, if the numbers were assigned to the third party administrator for each ERISA plan that they manage, then it sounded like they would have -- it would be like assigning a number to a Blue Cross plan with sub-numbers for each line of business, at least analogously. I was wondering if that was a correct assumption from what Mike just said.
MR. PAGELS: It is a very good question. One of the things that we are trying to capture is the relationship of that ERISA plan to that third party administrator. The question becomes, is it the employer that we want to give the number to, or is it the relationship with that third party administrator or that plan that needs to be enumerated.
There are some options there on how we go with that. One option would be to give the ERISA plan a number, and the other option would be to give each one of those plans associated with that ERISA, with that employer number. Does that make sense?
MR. GREENBERG: Yes.
MR. PAGELS: To be quite frank, it is an area that we are basically struggling with right now.
DR. LANDON: Rich Landon with Blue Cross/Blue Shield Association. A couple of the things that came out of the WEDI policy advisory group more than a couple of weeks ago, I guess the two most fundamental is that the payer I.D. should be used primarily for routing purposes, and second, it should have no embedded intelligence.
Now, relative to the second point, I think I heard that a consideration is being given to adding intelligence, in the sense of what I heard was plan type, and something else I wasn't clear on. Clarification: is that proposed for one of the digits of the identifier itself, or is it data elements of the supporting database?
MS. DIETRICK: The supporting database. It is not going to be within the number.
DR. LANDON: Thank you.
MR. PAGELS: One of the things that we should probably say, and I think it was in the handouts we gave you, in terms of embedded intelligence, I believe it was WEDI that came forward, and we were using 99 in the suffix to identify our medi-gap plans, and we did get rid of that.
DR. BRAITHWAITE: What Rich just said though brings up the point about the primary purpose of this identifier. From a provider's perspective, whether an ERISA plan is processed by TPA one or TPA two doesn't matter. The plan is identified as an ERISA plan and the TPAs might change over time. You don't want to have to change the identifier of that plan that the employee is enrolled in when that happens.
So you can change the routing information that is behind that number to a different TPA over time, but the number itself should not change.
DR. MOR: That's a good point. I think that is particularly important, since the third party administrators that are also insurers or payers risk certain product lines as well.
DR. COHN: Jeff, were you going to make a comment?
MR. BLAIR: I was just piggybacking on -- if you were going down the road that we seem to be recommending against, you would have to distinguish between the role of that insurance company, whether it is in the role of the TPA or whether it is in the role of a payer. I think we would want to avoid that.
DR. MOR: Given this routing notion then, the value of any information in the database begins to be quite limited for any purposes other than simply doing this administrative routing activity. I'm not sure the plan type makes a lot of difference.
DR. COHN: I'm not following you. Please explain.
DR. MOR: The value is in the I.D. system only, and not in the database. I guess that is what this discussion has just pointed out to me, that the value of any of the information in the database would be quite limited, or characterize the distribution of the insurance system in the U.S.
DR. COHN: Because there would be no historical data?
DR. MOR: Not only would there be no historical data, but it would be simply for routing purposes. You would have all these ERISA plans. You would have no understanding of what it was that would be -- the functions of the activities going on behind those. You wouldn't know what the distribution of managed care plans is in the country.
PARTICIPANT: I have a question about the routing. If you are going to merely identify that this is a Blue Cross, all you're doing -- Blue Cross is still going to require people who are submitting a claim to further identify it. They are not merely going to look at it and then try to figure out where you want to send it. They are going to give you a number or a code, so that they can automatically direct it within their system.
Why can't the provider I.D. have that to that kind of a detail, so that we are not then further required to put another number on it? I think personally that it needs to have down to the plan or at least down to the plan, or at least down to where did they want you to send this, so that when they are distributing within their company, they can properly send it for payment.
MS. DIETRICK: I think we addressed that earlier. That is a good possibility. We very well might do that, to the point that we are enumerating -- we have had feedback from both sides. We have had people tell us, just get it to the front door, be the gateway and we'll take it from there. Then we have had the other side of the coin, and now we need to hear industry tell us what you really, really need, which way we should go on this.
MS. GREENBERG: I wonder if we could have a little clarification of the WEDI recommendation that the primary purpose should be for routing, what was meant by that.
DR. LANDON: Again, I'm Rich Landon, wearing my WEDI hat this time. I will point out that I am not a WEDI board member. I can relay the information as it was discussed at the policy advisory group. This is not necessarily WEDI policy. So the board members from WEDI present, please correct me if I mis-speak.
One of the things that industry has been wrestling with is, what is the purpose of these identifiers. The policy advisory forum on the payer I.D. recommended or discussed their clear consensus, I guess, is a better word than recommendation, that the primary purpose of the payer I.D. be for routing, how to get a transaction, not just limited to claims, but all the HIPA transactions, the K2 transactions, to get it from the originator to the receiver, and that the primary purpose be exclusively for routing.
There are secondary purposes. The intelligence should be in the database. The previous question that was asked was very, very pertinent. The industry has a very, very clear position that it wants to be able to do it both ways.
Some companies, some payers who have opted to say we want one identifier either for the entire corporation or more probably for a given line of business, i.e., hospital claims. You route the claim, route the transaction to that address. We will take it internally and run with it.
Other speakers specifically referenced to the Veterans Administration say they have hundreds of sites. They want hundreds of different identifiers, so that each unique identifier will route that transaction to a unique site within the Veterans Administration.
I think the industry position is clear. We want it both ways and always in between, from a single I.D. to more than a hundred identifiers. I think what we have heard earlier is the earlier versions of the proposal that I have seen for payer I.D. limited to two suffixes.
The WEDI group discussed the option of moving it to what translates to three suffixes a thousand for identifier. The industry will continue to look at that. We are talking about a database that has not yet been developed, and I think we need to do some flow charting and modelling and testing before we really understand the impact of the current systems.
There was previous talk about, are we going to change identifiers, but it is also true for payer I.D. The probability is, most internal systems within payers will continue to use their unique identifiers. But the change will happen that, when the translation comes in the door, there will be a look-up table or a translator or a map or something to take from the NPI, the payer I.D., and map back to whatever identifier that particular application uses internally within the system.
Does that shed some light on your question?
MS. GREENBERG: Thank you.
DR. COHN: I guess I should make a comment, because I am on the board of WEDI, though I was not at the meetings where all this happened.
I would certainly, knowing the membership of WEDI, which is primarily a payer organization, not be surprised that the main focus in recommendations had to do with the very important aspect of the payer I.D. for routing and for payment of the bill.
It may be a part of the National Committee's duty to investigate this a little further to make sure that it also meets other secondary uses assuring quality of health care in the population and being able to identify down to that level. That may be an area where we can provide some value.
DR. DETMER: I think this has been very useful. It is apparent that this -- or at least, it seems to me, I am very appreciative of the comments, but it seems that this is not in some sense as far along, and that probably we will unfortunately get to crawl through the WEDIs instead of walk through the WEDIs, but at any event, hopefully not get too allergic about all of it.
I guess if there are ways that we can try to at least move the thing forward, that could be very useful. Clearly, I think the efficient routing is clearly a desirable and essential piece. The question is, how much further can we go, and should we try to go. I'm not sure we're going to be able to do that this morning. At least, that is what I'm hearing.
DR. COHN: Well, I guess I would ask the group, recognizing that we have heard a fair amount about the payer I.D., I think for this one, I would agree with you, it is hard to get the same level of confidence with the provider and yet, I would ask the group whether the sentiment is that this is a good idea, that the number itself seems to be a good number. They are going in the right direction, but we probably do need to have further hearings, or at least investigate this area a little further.
Thoughts from the committee?
DR. FRAWLEY: I would concur with your remarks, up to hearings. I think if there is a way that when we publish the notice in the Federal Register, to try to elicit more input from industry on possibly secondary uses, or where we think number has some utility other than as WEDI has envisioned it for routing.
But in terms of moving forward, I think we feel comfortable with the number, in terms of information, that we have been given. At some point, we're not going to have all of our questions answered. Obviously, there is still some staff work that is ongoing.
MS. GREENBERG: Can I ask a point of clarification?
DR. COHN: Please.
MS. GREENBERG: It would probably be interesting to watch the first wave, just give them one number and they all deal with it. Will however they deal with it then be in the database, or will however they have assigned numbers then be part of the payer I.D., or will it just be an unknown?
MR. PAGELS: In the situation where they were going to be a gateway, they would get one number and they would distribute it within their organization.
I don't know if there would really be a way for you to look at the database to understand that, but you would basically see that the organization has but one number. So you may make an assumption that in fact, that is how they are doing business as a gateway.
Does that make sense, Rich?
MS. GREENBERG: So like in the provider I.D. is an individual and an individual and an individual, here a payer is who knows, it sounds like. You really have apples and oranges.
MR. PAGELS: Again, any comments regarding how prescriptive we should be in this area would be greatly appreciated.
DR. COHN: Does the group have a sentiment?
DR. MOR: That is one of the reasons why we thought about this, listening to this. It has no research application. I think from that vantage point, it is just a number, which has some presumed administrative ease for routing or what have you, but it has no meaning whatsoever. And the component parts can't have any meaning, either, if it is as amorphous.
So I'm not even sure it makes any sense writing definitions for what the categories are, if the root information core is not going to be the same across the multiplicity of payers.
MS. GREENBERG: Right. Is this really in the spirit of what this unique payer I.D. was supposed to be, that it would be so diffuse? I don't know. I don't know enough about this area, I'd have to admit. But I guess I expected -- my anticipation was that it would be more of a consistent --
DR. MOORE: I think we started out with one goal, and then as we got industry input, we had our minds changed as to where we were going. One of the things that occurred at the session was, what is the purpose -- and this I stated back some months ago, this was an administrative simplification, not, are we doing data expansion.
What is the goal here? The goal here was to try to get electronic interactions, transactions, or electronic commerce from point A to point B, and make sure they are delivered to the right person. That is the primary goal.
DR. DETMER: And I'm not bothered by that. I think the point is, health care is changing so rapidly anyway, even if you did try to make these definitions, goodness knows whether there would be a creature the next year as you looked at it or not.
So I'm not that troubled. I think if you can get something going, that's fine.
MR. PAGELS: I just want to add to that, I guess the advantage is, when you get something going, then there is always the potential to add, if it seems reasonable.
DR. COHN: Right now, I am hearing two things simultaneously out of the committee, which is concern and lack of concern. If there is no concern, it is time for lunch. If there is concern, we will at least try to note it and figure out what to do next.
I guess the question is, is there concern or is there --
DR. MOR: Nothing that ruins my appetite, that's for sure.
DR. COHN: Okay. With that then, a final comment or question?
DR. LANDON: I did want to share one thing with the committee. This is the association's thoughts on both NPI and payer I.D. While we are real pleased with the progress that has narrowed the focus as to what the purpose of the identifiers are to the entire industry, with special thanks to the HCFA staff, we think we are getting a comfort level with the structure and format of the identifier itself.
We have not yet achieved a comfort level with how the supporting databases are going to work. I do specifically want to go on record with the committee as saying we have a concern with that, particularly because under the federal regulations, we are not privy to the draft of the NPRMs.
So until we see the NPRMs, we will not be able to do analysis and generate comments on that. Our concern is, since we are pushing the envelope with the sophistication of these databases, we are unsure of the purposes. Our concern is the 60-day comment period, and then HCFA's ability to understand the industry needs and requirements in getting from a proposed rule to the final rule may not be sufficient.
We would like not to make a specific recommendation on that. We are thinking that maybe we need a second NPRM rather than NPRM, then a final NPRM. But I do want to get that concern on the table, and to recap that. The concern is, we are not sure of how the supporting databases are going to work, and we are very, very uncomfortable that a comment period on the NPRM leading to a final rule will be adequate to insure smooth implementation under the HIPA time frames in a trillion dollar industry.
MR. PAGELS: Dr. Cohn, I would just like to add one final comment. Similar to what Karen did with the NPI, as we are trying to move forward with our NPRM, and if there are questions that this committee would present to us that it feels would strengthen our position or at least we could put forth in the NPRM, we would appreciate any feedback along those lines. So, thanks.
DR. COHN: I think the committee wants to thank both of you for joining us this morning. With that, why don't we adjourn until 1 o'clock for the afternoon session.
(The meeting adjourned for lunch at 12:15 p.m., to reconvene at 1:00 p.m.)