PDA

View Full Version : Actuarial Value Calculator is posted


Pages : [1] 2

thing
11-20-2012, 11:37 AM
http://cciio.cms.gov/resources/regulations/index.html

near the bottom of the page

http://cciio.cms.gov/resources/EHBBenchmark/av-calculator-final-locked-11-20-2012.xlsm

I can't get it to open. :curse:

cincinnatikid
11-20-2012, 12:11 PM
It's awesome...so many checkboxes...it's like xmas has come early!!!

thing
11-20-2012, 12:14 PM
post it to this site, if you can - still can't get the cciio version to download

thing
11-20-2012, 12:32 PM
got it. IT issue on my company's side, apparently.

IMP
11-20-2012, 12:34 PM
password?

plasticdog30
11-20-2012, 12:39 PM
The password is "password".

IMP
11-20-2012, 12:42 PM
i tried reef. :(

thing
11-20-2012, 12:43 PM
AABBABABAAAB

thing
11-20-2012, 01:51 PM
Good golly, it took me like 30 minutes to figure out how to override the VBA project password as well!

FormLetter
11-20-2012, 02:05 PM
how do you do that?

IMP
11-20-2012, 02:07 PM
magic

thing
11-20-2012, 02:25 PM
http://stackoverflow.com/questions/1026483/is-there-a-way-to-crack-the-password-on-an-excel-vba-project

It's fun, try it!

erosewater
11-20-2012, 03:54 PM
The regs are out as well. The rate review regulations cites the bogus study on medical bankruptcies I've railed about for years. HHS is straight up trolling me

FormLetter
11-20-2012, 03:55 PM
Also, it's possible to generate AVs above 100% in the AV calculator.

Has anyone been able to build a Bronze plan that has a deductible no higher than $2000?

thing
11-20-2012, 03:59 PM
I feel I should explain to anyone who's concerned about it - the entire guts of the AV Calculator are in VBA. The spreadsheet has numerous pop-ups built into it, for example when you select a certain cell. It takes the booleans for including coverage under a deductible etc. via a checkbox, instead of a 1/0 in a cell. In short, it is designed around the idea of a person sitting and manually entering the details of exactly one plan design into the template.

One of my first tasks is going to be to run the several hundred plan designs my company offers through, to get a sense of where currently the block is.

I see no way of doing this systematically and accurately without hacking the template.

Sorry, CMS/HHS/CCIIO. Excel is not, by nature, secure.

And as a service to the rest of ya'll, here it is unlocked:

erosewater
11-20-2012, 04:00 PM
awesome, thanks thing

ADoubleDot
11-20-2012, 04:14 PM
thanks thing :toast:

Maximus
11-20-2012, 04:47 PM
Also, it's possible to generate AVs above 100% in the AV calculator.

Has anyone been able to build a Bronze plan that has a deductible no higher than $2000?

Yes ... if you use a coinsurance percentage that is less than 50%. Not sure how well that goes over though.

Not Mike
11-20-2012, 04:48 PM
Why are there different continuance tables for platinum, gold, silver, and bronze? Or maybe I will discover this secret as I get into the model.

And, thanks for Unlocking!!

FormLetter
11-20-2012, 04:50 PM
Why are there different continuance tables for platinum, gold, silver, and bronze? Or maybe I will discover this secret as I get into the model.

And, thanks for Unlocking!!

Did you read the methodology document yet? I'm hoping it's answered there.

Not Mike
11-20-2012, 04:55 PM
Haven't read methodology yet. This is a pretty nice free tool, though.


Oh, and 6500 ded, with 6500 OOPM is a 57.5% AV.

My claim all along, re:affordability, is that it's nearly impossible to have a design that won't meet the 60% rule (I think back then I was using $6k).

thing
11-20-2012, 05:03 PM
bottom of page 1 of the pdf: "Because spending is affected by plan design through induced demand, the claims data is used to develop four sets of continuance tables..."

Not Mike
11-20-2012, 05:22 PM
bottom of page 1 of the pdf: "Because spending is affected by plan design through induced demand, the claims data is used to develop four sets of continuance tables..."

The problem that, in that case, you aren't measuring design. You are measuring a combination of design and utilization that is adjusted for the design.

Dismal Science
11-20-2012, 05:34 PM
The problem that, in that case, you aren't measuring design. You are measuring a combination of design and utilization that is adjusted for the design.

:iatp:

Seems like an odd choice

FormLetter
11-20-2012, 05:41 PM
bottom of page 1 of the pdf: "Because spending is affected by plan design through induced demand, the claims data is used to develop four sets of continuance tables..."

The problem that, in that case, you aren't measuring design. You are measuring a combination of design and utilization that is adjusted for the design.

Is that really a problem?

FormLetter
11-20-2012, 05:48 PM
.

Not Mike
11-20-2012, 06:01 PM
Is that really a problem?

Yes, if you are supposed to be measuring design, i.e., the value of the plan parameters. That should be independent of utilization. For example, if we are talking about HDHP, we would have two factors - design change and utilization adjustment, which are separate and distinct.

Both impact the cost, but the AV should measure the design, not the combination of both.

DoctorNo
11-20-2012, 07:23 PM
Some interesting stuff in the regs today - nice of them to put them out during Thanksgiving week.

Maximus
11-20-2012, 07:42 PM
Some interesting stuff in the regs today - nice of them to put them out during Thanksgiving week.

The fact that the comment period ends the day after Christmas is an especially nice touch (Apparently the 30 days begins on the date of publish in the federal register which will be November 26).

FormLetter
11-20-2012, 08:59 PM
Yes, if you are supposed to be measuring design, i.e., the value of the plan parameters. That should be independent of utilization. For example, if we are talking about HDHP, we would have two factors - design change and utilization adjustment, which are separate and distinct.

Both impact the cost, but the AV should measure the design, not the combination of both.

While we're at it, reimbursements are being entirely ignored.

Rainson
11-20-2012, 09:06 PM
I am a little disappointed. As one of the beta testers for the AV calculator I raised a concern about the inclusion of only individuals continuously enrolled for the entire 12 months. I see from a quick review of the instructions that this was not changed.

I pointed out the fact that deductibles are not prorated throughout the year (i.e. if you buy a $1,000 deductible plan effective July 1 you still have to meet a $1,000 deductible and not a $500 deductible) and the population included in the continuance table should reflect the average duration of the population you expect to insure. At first I thought this was just a significant issue for small group (i.e. people change employers throughout the year and this wouldn't be captured) and that in the individual market everyone would pretty much have 12 months of coverage as they would enroll during the open enrollment and very few would come in throughout the year due to a qualifying event. Then I considered the fact that people who are in great health may lapse coverage in the fourth quarter and take a chance, knowing they can get back in on January 1 of the next year. The fact that these individuals would also likely be the healthiest ones leading to premium reductions with little reduction in claims is also an issue that does not appear to be considered.

We have studied the average duration of those in small group during a given year among our clients and find that it is about 10 months, plus or minus. This means using a table with only people enrolled for 12 months will understate the value of the deductible and result in a calculated AV that is too high.

Thoughts?

Bamafan
11-20-2012, 09:52 PM
I am a little disappointed. As one of the beta testers for the AV calculator I raised a concern about the inclusion of only individuals continuously enrolled for the entire 12 months. I see from a quick review of the instructions that this was not changed.

I pointed out the fact that deductibles are not prorated throughout the year (i.e. if you buy a $1,000 deductible plan effective July 1 you still have to meet a $1,000 deductible and not a $500 deductible) and the population included in the continuance table should reflect the average duration of the population you expect to insure. At first I thought this was just a significant issue for small group (i.e. people change employers throughout the year and this wouldn't be captured) and that in the individual market everyone would pretty much have 12 months of coverage as they would enroll during the open enrollment and very few would come in throughout the year due to a qualifying event. Then I considered the fact that people who are in great health may lapse coverage in the fourth quarter and take a chance, knowing they can get back in on January 1 of the next year. The fact that these individuals would also likely be the healthiest ones leading to premium reductions with little reduction in claims is also an issue that does not appear to be considered.

We have studied the average duration of those in small group during a given year among our clients and find that it is about 10 months, plus or minus. This means using a table with only people enrolled for 12 months will understate the value of the deductible and result in a calculated AV that is too high.

Thoughts?

Seems like an issue that will not favor any one issuer over another. Nothing seems to prevent me from pricing for the true expected value of the plan benefits, relative to the index rate. The rating rules explicitly limit plan rating variations from the index rate to AV/cost-sharing design, network/UM differences, benefits in addition to EHB's, and catastrophic plan eligibility provisions.

Bamafan
11-20-2012, 10:26 PM
I feel I should explain to anyone who's concerned about it - the entire guts of the AV Calculator are in VBA. The spreadsheet has numerous pop-ups built into it, for example when you select a certain cell. It takes the booleans for including coverage under a deductible etc. via a checkbox, instead of a 1/0 in a cell. In short, it is designed around the idea of a person sitting and manually entering the details of exactly one plan design into the template.

One of my first tasks is going to be to run the several hundred plan designs my company offers through, to get a sense of where currently the block is.

I see no way of doing this systematically and accurately without hacking the template.

Sorry, CMS/HHS/CCIIO. Excel is not, by nature, secure.

And as a service to the rest of ya'll, here it is unlocked:

Whoa. There's over 5,500 lines of VBA code in Module1.


Thanks for the unlocked version! Will you be providing a version that will run plans in batch as well? :wink:

thing
11-20-2012, 11:16 PM
Thanks for the unlocked version! Will you be providing a version that will run plans in batch as well? :wink:
Hmmm... I was too quick to post that, wasn't I? Could have put together a excel formula version and sold it for $14K/company... ;-)

thing
11-20-2012, 11:26 PM
LIve-blogging as I read through the pdf.

On the underlying data: they start with a set of 54 million people. The narrative says they end with 12.5 million enrollees. But the continuance table for Platnum shows 1.4 million, Gold 3.9, Silver 0.4 and Bronze 0.35 million lives. So that doesn't add up too well.

Also, a bit concerned about dividing into age/sex/plan value cells, trending, and recombining based on some hypothetical age/sex distribution for each metal tier. Do they think Gold is going to have 10 times the enrollment of Silver? Is that what the membership on each table means? Probably reading too much into that.

thing
11-20-2012, 11:46 PM
No additional adjustment applied for "habilitative" services. Notes they are "transitional". This seems wrong and possibly dangerous. I can forsee this being designated as a pretty expensive category if the right interest groups get ahold of it.

thing
11-21-2012, 12:07 AM
Spent a little time looking at the VBA. Boy could they use a "case" statement.

I might just hack it to run against all 4 metal tiers and see how much variance is caused by the metal tier assumption...

thing
11-21-2012, 12:22 AM
12,000 checkboxes and dropdowns, and you can't say to not have office visit copays apply to the MOOP?

"Small differences between the modified MOOP calculated by this method and the exact MOOP that applies are unlikely to have a significant effect..."

Oh, that's alright then!

thing
11-21-2012, 12:35 AM
Overall:

How they derived the continuance tables is a little wacky
How they (didn't) adjust for different populations to be rated is a lot wacky
How they combined data for pediatric dental/vision is probably about right
How they didn't adjust for habilitative services is a time-bomb
How they implemented the calculator to throw up msgboxes and so forth is a pain
They brush over a lot of issues that were presumably brought by the beta testers

I'm nervous.

T-roy Boy
11-21-2012, 08:57 AM
The regs are out as well. The rate review regulations cites the bogus study on medical bankruptcies I've railed about for years. HHS is straight up trolling me

Prolly could use another thread for these. I was surprised that they are allowing multiple dependents rather then the normal 4 tier (EE, E+SP, EE+DEP, EE+FAM) policies. These are "preliminary" tho right? We have what, 30 days to really see what is going to stick.

exponentialpi
11-21-2012, 11:27 AM
Prolly could use another thread for these. I was surprised that they are allowing multiple dependents rather then the normal 4 tier (EE, E+SP, EE+DEP, EE+FAM) policies. These are "preliminary" tho right? We have what, 30 days to really see what is going to stick.

Based upon the proposed rules, they are also proposing a cap of rating only for the 3 oldest family members under the age of 21.

DoctorNo
11-21-2012, 12:22 PM
I'm intrigued by the prescribed age rating curve.

Zee
11-21-2012, 01:16 PM
Prolly could use another thread for these. I was surprised that they are allowing multiple dependents rather then the normal 4 tier (EE, E+SP, EE+DEP, EE+FAM) policies. These are "preliminary" tho right? We have what, 30 days to really see what is going to stick.

30 days from 11/26 to comment, so due by 12/26. Comments are reviewed/considered, & answers are developed in conjunction with the prep of Final Rules.

We are almost certainly looking at late January for final guidance, I think. That should leave PLENTY of time for final rate development, QHP cert and other rate filing prep... :wink:

DoctorNo
11-21-2012, 01:34 PM
The "single index rule" also concerns me.

EDIT: Actually, now that I've thought about it some more, it's not that bad.

ahensen
11-21-2012, 03:20 PM
So is this only applicable to small employer and individual? I recall something about there was going to be two calculators released, an Actuarial Value and a Minimum Value. The minimum value was suppose to be for self funded plans.

ahensen
11-21-2012, 03:22 PM
Also, is the OOP max excluding deductible or including? I can't find guidance.

thing
11-21-2012, 03:50 PM
Also, is the OOP max excluding deductible or including? I can't find guidance.

Includes.

Page 16: "Next, the calculator computes the spending level at which the modified MOOP will apply. To do so, the calculator subtracts the deductible from the modified MOOP and divides the resulting value by one minus the coinsurance rate,..."

erosewater
11-21-2012, 04:07 PM
Setting the deuctible and OOP max to 6500 gives an error that the deductible is higher than the OOP max. Same if I lower it to 6400. Weird

FormLetter
11-21-2012, 04:08 PM
What is weird about that? 6400 > 6500

Rainson
11-21-2012, 07:13 PM
So is this only applicable to small employer and individual? I recall something about there was going to be two calculators released, an Actuarial Value and a Minimum Value. The minimum value was suppose to be for self funded plans.

Yes, there is supposed to be an MV calculator as well. I'm not sure how different it will look, other than perhaps different underlying continuance tables.

Rainson
11-21-2012, 07:19 PM
Prolly could use another thread for these. I was surprised that they are allowing multiple dependents rather then the normal 4 tier (EE, E+SP, EE+DEP, EE+FAM) policies. These are "preliminary" tho right? We have what, 30 days to really see what is going to stick.

I'm pretty sure the member rating approach will stick, at least in the individual market. They need this for a few reasons:

1. By law the tobacco loads can only be applied to the portion of the premium for those members of the family that use tobacco.
2. In the individual market they need each member to have a premium assigned to them so that they have some way of allocating the premium subsidy across carriers when individual family members select different carriers.

thing
11-23-2012, 03:07 PM
It's awesome...so many checkboxes...it's like xmas has come early!!!

190 checkboxes, dropdowns, and cells you can fill out.

Additionally, there are 24 checkboxes you can't fill out, but which are still on the sheet for some reason. 4 each for:
1. pediatric vision
2. pediatric dental
3. "abu", which I think is for substance abuse OP
4. physical therapy
5. well baby
6. Maintenance Rx drugs

Originally, I was thinking I would de-VBA-ify the spreadsheet, but that seems impractical. I'm instead making minor edits to their code (pretty much just commenting out all msgbox lines), creating a table of 189 variables, writing a macro to push those to the AV Calculator tab, then running Button 108 macro for each metal tier and storing the output.

I might be done by Sunday.

Zee
11-24-2012, 11:35 AM
I'm pretty sure the member rating approach will stick, at least in the individual market. They need this for a few reasons:

1. By law the tobacco loads can only be applied to the portion of the premium for those members of the family that use tobacco.
2. In the individual market they need each member to have a premium assigned to them so that they have some way of allocating the premium subsidy across carriers when individual family members select different carriers.

Which is why the 3-members-under-21 cap doesn't make any sense.

Also, I think risk adjustment is simplified, and perhaps the estimation of cost-sharing reduction too (i.e. the amount that will be subsidized to carriers).

thing
11-24-2012, 04:08 PM
3-members-under-21 cap will work for the majority of families to split the tobacco load. But it doesn't eliminate the problem - companies will still need a policy to deal with the 4 child family where exactly 1 child smokes. It'd be better to allow some kind of second-hand smoke rating rule, e.g. you can rate up to +50% for an individual smoker policy and up to +20% for the whole family where some but not all the covered members smoke.

Got my first bunch of plans run through the AV Calculator. No big surprises yet. Vast majority of my members have coverage outside the +/-2% of one of the metal tiers.

thing
11-24-2012, 07:14 PM
Setting the deuctible and OOP max to 6500 gives an error that the deductible is higher than the OOP max. Same if I lower it to 6400. Weird

Very weird. We've got a HDHP with a 100% coverage after the deductible. It gets the same error message. Checkbox_Int checked. Seems to need $100 difference between deductible and OOP, probably because they won't really use what you say is the OOP. The adjustment for preventive care, then, is $100?

Anyway, so far I'm getting around it by adding $100 to the declared OOP.

thing
11-25-2012, 03:22 PM
Maybe I am misunderstanding something.

If you look at row 88 of any of the continuance tables, I believe that is supposed to represent the total annual costs for all members for that coverage level for that benefit. For example, Gold Rx shows "Avg. Cost per Enrollee (Max'd)" of $1,099.91. However, it also shows Generic/Brand/Non-Pref/Specialty of $740.35/$151.43/$67.81/$114.66, which sums to $1,074.25 - a difference from the total of $25.66. In fact, the sum of the parts for all the Rx continuance tables is short of the total for each Rx continuance table by $25.66. This seems like a possible error in the model.

Likewise, when I attempt to sum the parts of the Medical table, I come up short of the total by $265 & change. The combined tables end up short of the total by about $291, except for Platinum, which is short by $304.38.

For some reason the Platinum combined table says SNF is $0 where the Platinum medical table says SNF is $7.32, which if you correct for that the Platinum combined is also off by about $291 from the parts.

Now, on the medical & combined table I am willing to believe that the discrepancy has something to do with Pediatric Vision and Dental, which would have been accounted for in the total but not in the parts. Aside from the likely mistreatment of Platinum Combined SNF, that seems like a plausible explanation. But that should not have had an impact on the Rx tables.

afg5008
11-26-2012, 08:43 AM
It looks like the calculator doesn't have a copay option for Outpatient Surgery. How do you guys plan on dealing with this? I didn't see anything in the regulation addressing this. I understand that the data they have doesn't support copays for OP surgery, but you'd think there'd be some sort of standard guidance.

FormLetter
11-26-2012, 09:42 AM
My understanding is that you can use the HHS calculator to get a measure of Actuarial Value to demonstrate that any particular plan design qualifies as one of the four metal levels based on only the aspects of the benefit design that the HHS AV calculator can accomodate.

So if you have two plan designs that are identical except for the OP surgery copay, then they would have the same AV and be the same metal level, despite having differences that could be material.

Bamafan
11-26-2012, 09:48 AM
My understanding is that you can use the HHS calculator to get a measure of Actuarial Value to demonstrate that any particular plan design qualifies as one of the four metal levels based on only the aspects of the benefit design that the HHS AV calculator can accomodate.

So if you have two plan designs that are identical except for the OP surgery copay, then they would have the same AV and be the same metal level, despite having differences that could be material.

I have a different understanding of the proposed rules. I thought I read that for any material differences between the benefit design and what you can get into the AV model, you can either make actuarially equivalent inputs to the model or you can make adjustments to the AV produced at the end. In either case, you have to also provide actuarial certification of the results by a member in good standing with the Academy.

FormLetter
11-26-2012, 10:22 AM
So if you decide not to make any adjustments to the output of the HHS AV calculator, do you need to provide actuarial certification that you determined they weren't material?

thing
11-26-2012, 11:54 AM
That's my read, as well. And yes, if you didn't adjust for something, you would state it wasn't material.

I've been putting in equivalent values for certain categories where the AV calculator doesn't make distinctions that my plan design makes, e.g. for incentive generic drugs. (Not sure that's a standard term - I mean generics on the Target/Walmart/etc. $4 list.)

Reik483
11-26-2012, 11:58 AM
Does the calculator have the ability to calculate the AV for a plan that has no drug coverage? If I de-select the box for the drugs to be subject to deductible and coinsurance and leave the "coinsurance, if different" and "copay, if different" fields blank does it just assume there's no coverage for those?

I'm trying to calculate the AV for ASO clients that carve out their drug coverage.

Dismal Science
11-26-2012, 12:24 PM
Does the calculator have the ability to calculate the AV for a plan that has no drug coverage? If I de-select the box for the drugs to be subject to deductible and coinsurance and leave the "coinsurance, if different" and "copay, if different" fields blank does it just assume there's no coverage for those?

I'm trying to calculate the AV for ASO clients that carve out their drug coverage.

By "carve out drug coverage" I assume you mean they use an independent Pharmacy Benefit Manager (PBM)?

If that is the case, to assess whether a plan meets minimum ARV threshold you would still include the pharmacy benefit bundled with the medical coverage - even if it is administered by another company.

afg5008
11-26-2012, 01:33 PM
190 checkboxes, dropdowns, and cells you can fill out.

Additionally, there are 24 checkboxes you can't fill out, but which are still on the sheet for some reason. 4 each for:
1. pediatric vision
2. pediatric dental
3. "abu", which I think is for substance abuse OP
4. physical therapy
5. well baby
6. Maintenance Rx drugs

Originally, I was thinking I would de-VBA-ify the spreadsheet, but that seems impractical. I'm instead making minor edits to their code (pretty much just commenting out all msgbox lines), creating a table of 189 variables, writing a macro to push those to the AV Calculator tab, then running Button 108 macro for each metal tier and storing the output.

I might be done by Sunday.

I found 1-5, but I don't see any reference to the maintenance drugs. What is the name for the checkbox they are using for that one? Also, for 1-5 there are hidden, blank columns in the continuance tables for those items.

Edit:
Nevermind. It shows up in "ThisWorkbook" but not in Module 1 where I was looking before.

GhostMac
11-27-2012, 10:48 AM
Anybody try running 2 tier plans? I am getting a runtime error 6 when attempting to do so.

My Dogs Are Barking!
11-27-2012, 01:53 PM
Yes, I had the same issue.

GhostMac
11-27-2012, 03:38 PM
MDAB, do you know of a workaround?

afg5008
11-27-2012, 03:57 PM
Note that their macro looks up the wrong row for OT/PT. The tool ignores any cost sharing entered into row 29. Instead, you need to enter it in the hidden and greyed out row 28.

thing
11-27-2012, 04:40 PM
Note that their macro looks up the wrong row for OT/PT. The tool ignores any cost sharing entered into row 29. Instead, you need to enter it in the hidden and greyed out row 28.

That's problematic. I assume the error would also mean it's looking at the (non-used) ded_pt and coins_pt checkboxes as well.

GhostMac
11-27-2012, 04:59 PM
The issue with a 2 tier plan design is if you enter Rx copays into the 2nd tier section, the program crashes. Are we not supposed to enter Rx copays into tier 2?

Sir Deicide
11-27-2012, 06:22 PM
Out of network costs do no count in the AV calculation. Does anyone know how indemnity plans are to be treated?

thing
11-27-2012, 06:30 PM
Setting the deuctible and OOP max to 6500 gives an error that the deductible is higher than the OOP max. Same if I lower it to 6400. Weird

Found it. The following block of code is the culprit:

If Sheet1.CheckBoxes("Checkbox_Int").Value = 1 Or Sheet1.CheckBoxes("Checkbox_Sep").Value <> 1 Then
If (deduct + deduct_rx > oop) Then
Sheet1.Range("B55").Value = "Error: Deductible value is higher then maximum out of pocket value."
Sheet1.Range("B56").Value = ""
Sheet1.Range("B57").Value = ""
Exit Sub
End If


When Checkbox_Int = 1, it doesn't set the variable deduct_rx, meaning if that variable was set in a previous run of the macro, it retains that value. Which can lead to errors. Obviously.

thing
11-27-2012, 06:34 PM
By the way, I have now been in contact with HHS regarding the issues I have uncovered and the other issues reported here. I'll keep ya'll posted how that goes.

DoctorNo
11-27-2012, 08:59 PM
Quick silly question (and I'm not in the weeds doing these calculations myself, so forgive any potential silliness):

Cells E35/E36 (for OP facility/surgery copays) are grayed out, signifying that they're not to be used. However, when pricing a two-tier plan, the corresponding cells for tier 2 (I35/I36) are *not* grayed out.

Can you actually put values in here and have the calculations work (and therefore, the graying out is a flaw), or can you really only put values in for tier 2?

afg5008
11-28-2012, 09:24 AM
That's problematic. I assume the error would also mean it's looking at the (non-used) ded_pt and coins_pt checkboxes as well.

No, it looks like it's using the correct checkboxes.



Thanks for contacting HHS. I look forward to your updates.

afg5008
11-28-2012, 09:58 AM
I’ve entered plans that are all copays except for OP facility/professional, which I give a custom coinsurance. I would expect the general deductible and coinsurance to have zero effect on the AV for these plan designs; but changing them does affect the AV. Are the general deductible and coins being applied to things not shown on the “AV Calculator” tab (perhaps pediatric dental/vision)? I’m not sure how to interpret these results.

accolade82
11-28-2012, 10:40 AM
Does anyone else find this to be an issue:

Enter an integrated $2,500 Deductible, 100% Coins, with $2,500 OOP max, all services subject to deductible and coins (except preventive care), and vary the Assumed metal tier. I get the following results:

Platinum: 75.0%
Gold: 75.2%
Silver: 74.9%
Bronze: 75.6%

The documentation indicates that different continuance tables reflect different degrees of demand, which make sense to me: Better benefits implies higher utilization.

I know this isn't how the tool was intended to be used, but given the methodology adopted, I would have expected the Actuarial Values in the above example to decrease as the metals moved from platinum to bronze, not that realized chaotic pattern above. It doesn't make sense to me that a $2,500 deductible using a continuance table with low utilization (bronze) would have an AV higher than a $2,500 deductible using a continuance table with high utilization (platinum).

Someone please enlighten me if I am wrong.

jayhawk
11-28-2012, 10:55 AM
Maybe I am misunderstanding something.

If you look at row 88 of any of the continuance tables, I believe that is supposed to represent the total annual costs for all members for that coverage level for that benefit. For example, Gold Rx shows "Avg. Cost per Enrollee (Max'd)" of $1,099.91. However, it also shows Generic/Brand/Non-Pref/Specialty of $740.35/$151.43/$67.81/$114.66, which sums to $1,074.25 - a difference from the total of $25.66. In fact, the sum of the parts for all the Rx continuance tables is short of the total for each Rx continuance table by $25.66. This seems like a possible error in the model.



Is it possible that one of the continuance tables refer to utilization and one to cost? 75% of rx cost should not be generic drugs.

afg5008
11-28-2012, 11:30 AM
Does anyone else find this to be an issue:

Enter an integrated $2,500 Deductible, 100% Coins, with $2,500 OOP max, all services subject to deductible and coins (except preventive care), and vary the Assumed metal tier. I get the following results:

Platinum: 75.0%
Gold: 75.2%
Silver: 74.9%
Bronze: 75.6%

The documentation indicates that different continuance tables reflect different degrees of demand, which make sense to me: Better benefits implies higher utilization.

I know this isn't how the tool was intended to be used, but given the methodology adopted, I would have expected the Actuarial Values in the above example to decrease as the metals moved from platinum to bronze, not that realized chaotic pattern above. It doesn't make sense to me that a $2,500 deductible using a continuance table with low utilization (bronze) would have an AV higher than a $2,500 deductible using a continuance table with high utilization (platinum).

Someone please enlighten me if I am wrong.

My guess is that it's the bolded that is causing the issue.

thing
11-28-2012, 12:53 PM
No, it looks like it's using the correct checkboxes.

:tup:
Thanks for reassuring me on that point!

thing
11-28-2012, 12:57 PM
Quick silly question (and I'm not in the weeds doing these calculations myself, so forgive any potential silliness):

Cells E35/E36 (for OP facility/surgery copays) are grayed out, signifying that they're not to be used. However, when pricing a two-tier plan, the corresponding cells for tier 2 (I35/I36) are *not* grayed out.

Can you actually put values in here and have the calculations work (and therefore, the graying out is a flaw), or can you really only put values in for tier 2?

On the locked version, the copay for OP fac/prof for Tier 2 are locked cells, and you can't even select them. Failure to gray out is just (another) coding sloppiness moment.

thing
11-28-2012, 12:58 PM
Are the general deductible and coins being applied to things not shown on the “AV Calculator” tab (perhaps pediatric dental/vision)?
I believe this to be true, but I am not certain.

thing
11-28-2012, 01:17 PM
The documentation indicates that different continuance tables reflect different degrees of demand, which make sense to me: Better benefits implies higher utilization.
I think they got this backwards, actually. My reading is that the continuance tables are for the plan's paid amounts, meaning you should divide the Platinum by 0.9 etc. to get the member's contract allowed amounts. Doing so gives you higher amounts for Bronze than for Platinum, sort of like a non-managed versus managed assumption. Which is ridiculous.

I recall there being a note about utilization assumptions being in the Platinum and Gold, but not the Silver nor Bronze. Am I making that up? Not finding it now.

Not Mike
11-28-2012, 01:26 PM
Does anyone else find this to be an issue:

Enter an integrated $2,500 Deductible, 100% Coins, with $2,500 OOP max, all services subject to deductible and coins (except preventive care), and vary the Assumed metal tier. I get the following results:

Platinum: 75.0%
Gold: 75.2%
Silver: 74.9%
Bronze: 75.6%

The documentation indicates that different continuance tables reflect different degrees of demand, which make sense to me: Better benefits implies higher utilization.

I know this isn't how the tool was intended to be used, but given the methodology adopted, I would have expected the Actuarial Values in the above example to decrease as the metals moved from platinum to bronze, not that realized chaotic pattern above. It doesn't make sense to me that a $2,500 deductible using a continuance table with low utilization (bronze) would have an AV higher than a $2,500 deductible using a continuance table with high utilization (platinum).

Someone please enlighten me if I am wrong.

The pattern seems odd. I'd expect the AVs to INCREASE (from platinum to bronze)if they are building utilization assumptions in, which I'll reiterate is inappropriate when assessing plan value.

Why increase? Well, because utilization of "discretionary" services should decrease, which means that less people are using care at lower dollar amounts, which means that the care received is more highly skewed to large dollar (unavoidable) claims, which have higher plan cost sharing.

But at the end of the day, we should be assessing overall benefit value. And that should be assessed on the ENTIRE population of people, not on a slice.

FormLetter
11-28-2012, 01:46 PM
I’ve entered plans that are all copays except for OP facility/professional, which I give a custom coinsurance. I would expect the general deductible and coinsurance to have zero effect on the AV for these plan designs; but changing them does affect the AV. Are the general deductible and coins being applied to things not shown on the “AV Calculator” tab (perhaps pediatric dental/vision)? I’m not sure how to interpret these results.

This sounds like a problem to me.

accolade82
11-28-2012, 03:04 PM
The pattern seems odd. I'd expect the AVs to INCREASE (from platinum to bronze)if they are building utilization assumptions in, which I'll reiterate is inappropriate when assessing plan value.

Why increase? Well, because utilization of "discretionary" services should decrease, which means that less people are using care at lower dollar amounts, which means that the care received is more highly skewed to large dollar (unavoidable) claims, which have higher plan cost sharing.

But at the end of the day, we should be assessing overall benefit value. And that should be assessed on the ENTIRE population of people, not on a slice.

Yeah, that makes sense for higher deductibles.

thing
11-28-2012, 03:31 PM
Is it possible that one of the continuance tables refer to utilization and one to cost? 75% of rx cost should not be generic drugs.
There are separate columns built in for utilization and cost.

FractionalActuary
11-29-2012, 09:03 AM
I cannot stop hitting errors when trying to automate steps. No matter what I do I keep getting values of multiply in the VBA that correspond to columns containing "." in the continuance tables. It is normally due to hidden boxes being checked, but now I am even getting it due to the specialist coinsurance being checked, which should be standard.

Anyone else having similar issues?

GhostMac
11-29-2012, 09:17 AM
I am getting an error when inserting a "coinsurance if different" amount in the "specialist vist" category when the plan is run as nonintegrated med/rx. That error goes away when the plan is run as integrated med/rx, but it will error out if you enter a "coinsurance if different" amount in the "Rehabilitative Occupational and Rehabilitative Physical Therapy" category.

afg5008
11-29-2012, 11:22 AM
This sounds like a problem to me.

I think it has something to do with the way the OOPM is being used. If I set all cost shares to 0% and set the general coins/OOPM to 10%/$6,000, then the calculator makes an adjustment by adding to the numerator for claims made above 6,000/.9 = 6,667. But if I change the general coins to 90%, then that actually decreases my AV since it only adds to the numerator for claims over $6,000/.1 = $60,000.

Reik483
11-29-2012, 11:52 AM
I was messing around with the calculator and setting a generic copay to 2000 brings back an AV of ~700%. Anyone else having issues with AVs above 100%?

afg5008
11-29-2012, 01:27 PM
I was messing around with the calculator and setting a generic copay to 2000 brings back an AV of ~700%. Anyone else having issues with AVs above 100%?

What did the rest of your cost sharing look like? I haven't gotten anything above 100%.

Reik483
11-29-2012, 05:37 PM
I managed to get an AV of like 360% and one of -133.7% (so l33t)

http://imgur.com/a/w4XCp

thing
11-29-2012, 06:33 PM
I managed to get an AV of like 360% and one of -133.7% (so l33t)

http://imgur.com/a/w4XCp

Awesome.

thing
11-29-2012, 07:00 PM
You can 'fix' that by selecting the lesser_of checkbox, by the way.

thing
11-29-2012, 07:41 PM
On the locked version, the copay for OP fac/prof for Tier 2 are locked cells, and you can't even select them. Failure to gray out is just (another) coding sloppiness moment.

Also, if you comment out the block that pops up the error message, you find that the value inside these cells does impact the AV.

thing
11-29-2012, 07:52 PM
Note that their macro looks up the wrong row for OT/PT. The tool ignores any cost sharing entered into row 29. Instead, you need to enter it in the hidden and greyed out row 28.
Correction - it does look at the value in row 29, for coinsurance at least. It just then references to one of the hidden and blank columns in the continuance tables, and returns a run-time error (type mismatch) because "." is not a number.

jayhawk
11-29-2012, 10:27 PM
There are separate columns built in for utilization and cost.

So the tables have a GDR of about 91% and an average cost per generic drug over $75. I hope the costs in the tables are billed charges? It has 75% of the total drug costs as generic drugs.

FormLetter
11-30-2012, 07:17 AM
You can 'fix' that by selecting the lesser_of checkbox, by the way.

It is terrible design that you have to select an option to make the results be plausible, if not credible.

db20
11-30-2012, 09:10 AM
Does anyone else find this to be an issue:

Enter an integrated $2,500 Deductible, 100% Coins, with $2,500 OOP max, all services subject to deductible and coins (except preventive care), and vary the Assumed metal tier. I get the following results:

Platinum: 75.0%
Gold: 75.2%
Silver: 74.9%
Bronze: 75.6%


Does this pattern reflect the subsidies kicking in at the silver level?

FormLetter
11-30-2012, 09:48 AM
Which subsidies are you referring to? The cost-sharing subsidies?

db20
11-30-2012, 10:28 AM
Which subsidies are you referring to? The cost-sharing subsidies?

Yes, specifically looking at the table on the 2nd page here:

http://www.kff.org/healthreform/upload/8177.pdf

Still just speculation, but it looks like the pattern might have gone 75.6/ 75.4/ 75.2/ 75.0 if the AVs of subsidized plans falling between 68% and 72% would have instead fallen outside of the silver with de minimis variation. It looks like this only occurs for silver plans since the other plans based on the FPL are 73%, 87%, and 94% which will not qualify as metallic plans, thus not manifesting in the AV calculator.

thing
12-03-2012, 07:44 PM
If you take the cost column divided by the freq column, this is intended to represent the cost/service for people below that deductible level (or something). That is how the VBA determines the 'lesser of' when you check that boolean, for example. For most of the services, this produces about what you'd expect, a monotonically increasing column leveling out at about what you'd normally have as allowed costs for that service category. But, in a couple of cases, you get whacked out, rising, then falling, then rising again patterns.

In particular, the IP column is nuts. Below is the first part of that calculation for IP, for each of the metal levels:


Up To P G S B
$0 0.00 0.00 0.00 #DIV/0!
$100 389.03 2,909.98 794.71 #DIV/0!
$200 3,757.35 7,717.88 1,677.01 #DIV/0!
$300 7,893.40 9,259.98 3,865.43 5,411.76
$400 18,064.48 11,890.50 6,413.69 6,684.15
$500 27,549.20 10,053.73 7,108.06 10,641.26
$600 21,770.06 10,407.47 11,494.78 15,374.24
$700 19,683.77 9,940.36 13,797.54 12,186.47
$800 16,583.59 8,727.31 13,009.16 9,702.20
$900 19,509.69 8,245.62 11,145.81 8,577.28
$1,000 18,292.63 7,772.59 9,152.43 6,940.11
$1,100 16,164.25 7,347.82 7,145.50 5,686.70
$1,200 15,653.85 6,637.32 6,857.44 4,749.06
$1,300 14,460.67 6,039.98 5,640.57 4,389.69
$1,400 12,202.76 5,458.16 5,087.63 3,653.29
$1,500 10,660.07 5,040.96 4,352.95 3,253.32
$1,600 9,411.43 4,571.42 3,836.47 3,014.24
$1,700 8,232.44 4,249.89 3,670.86 2,847.27
$1,800 7,420.38 4,062.91 3,400.77 2,704.94
$1,900 6,657.64 3,771.47 3,223.74 2,618.87
$2,000 6,256.80 3,551.92 3,033.89 2,500.15


I've noticed similar but less extreme version in speech therapy, there are probably others, I haven't checked them all yet. Sending another note to CMS. Thought ya'll should be aware.

Dismal Science
12-03-2012, 08:19 PM
Has CMS responded to any of your comments yet?

thing
12-03-2012, 10:14 PM
Wow, that formatting really failed.
@dismal, nothing in writing.

DoctorNo
12-03-2012, 10:20 PM
I imagine that they're going to hold off on commenting until after the comment period expires on the 26th.

FormLetter
12-04-2012, 07:05 AM
I wonder who built this thing.

DoctorNo
12-04-2012, 09:44 AM
I wonder who built this thing.

Per the documentation that comes with the model, they're a firm that doesn't employ actuaries.

thing
12-04-2012, 11:45 AM
Blue Health Intelligence (https://www.bluehealthintelligence.com/)

Likely had some actuarial input - their 'career' section takes you to job openings at individual Blues carriers, though.

DoctorNo
12-04-2012, 12:09 PM
Blue Health Intelligence (https://www.bluehealthintelligence.com/)

Likely had some actuarial input - their 'career' section takes you to job openings at individual Blues carriers, though.

That's where the data came from. The work was done by Acumen in San Francisco, led by two PhDs.

thing
12-04-2012, 02:07 PM
Ah.

So, I've been reading the VBA. This is not a healthy pursuit. My conclusions so far:

1. everywhere the code has a line like

numerator = numerator {+, -} ded_freq * ((ded_amt) / (ded_freq) - copay_amt)

it ought to read

numerator = numerator {+, -} ded_freq * Application.WorksheetFunction.Max(((ded_amt) / (ded_freq) - copay_amt), 0)


The reason for this formula is to account for the amount of a service beyond what is covered by the copay. However, if you enter a copay greater than the average cost of the service, it treats it as a negative cost to the insurer. You can create big exceptions by doing this to the Rx portion, since the Rx portion is relatively low cost, but similar errors are 'buried' in the AV if this trick is done on, say, office visits with a 'copay' of over $130 or so. Since, in general, a doctor's office will not treat (the very rare) situations where the copay is greater than the office charge as an obligation to write the insurer a check, this formula is not representative of what really happens with cost sharing.

2. The code always checks for the impact of the coinsurance for costs between the deductible and the MOOP (both as modified by copay lines). Even if, as is common for Rx, none of the benefits are subject to the deductible or coinsurance. This means a straight copay medical plan has the same problem in spades. Those blocks of code need to be wrapped in an "if...then...." block that checks that any of the referenced benefits are in fact subject to deductible/coinsurance, or errors result. The honest workaround is to set the coinsurance level to 0% in these cases. Intuitively, though, you would set it 100%, and that does result in a AV that can be nearly 10% higher. And for your entertainment, there is not a restriction that the coinsurance be between 0% and 100%, so this bug can be used to get AV below 0% and above 100%.

Yo Momma
12-04-2012, 03:32 PM
So...has anyone coded this to batch rate multiple benefit designs? If so, care to share so that I can be saved the time to do this?

thing
12-04-2012, 03:55 PM
So...has anyone coded this to batch rate multiple benefit designs?Yes. Yes, I have. If so, care to share so that I can be saved the time to do this?I'm not sure that's reasonable. Aside from the doing-work-for-the-competition-and-moreover-for-free aspect of it, I'm still auditing. Turns out the code doesn't automatically reset every variable it ought to when you run it in a batch mode.

afg5008
12-04-2012, 04:31 PM
Ah.

So, I've been reading the VBA. This is not a healthy pursuit. My conclusions so far:

1. everywhere the code has a line like

numerator = numerator {+, -} ded_freq * ((ded_amt) / (ded_freq) - copay_amt)

it ought to read

numerator = numerator {+, -} ded_freq * Application.WorksheetFunction.Max(((ded_amt) / (ded_freq) - copay_amt), 0)


The reason for this formula is to account for the amount of a service beyond what is covered by the copay. However, if you enter a copay greater than the average cost of the service, it treats it as a negative cost to the insurer. You can create big exceptions by doing this to the Rx portion, since the Rx portion is relatively low cost, but similar errors are 'buried' in the AV if this trick is done on, say, office visits with a 'copay' of over $130 or so. Since, in general, a doctor's office will not treat (the very rare) situations where the copay is greater than the office charge as an obligation to write the insurer a check, this formula is not representative of what really happens with cost sharing.

2. The code always checks for the impact of the coinsurance for costs between the deductible and the MOOP (both as modified by copay lines). Even if, as is common for Rx, none of the benefits are subject to the deductible or coinsurance. This means a straight copay medical plan has the same problem in spades. Those blocks of code need to be wrapped in an "if...then...." block that checks that any of the referenced benefits are in fact subject to deductible/coinsurance, or errors result. The honest workaround is to set the coinsurance level to 0% in these cases. Intuitively, though, you would set it 100%, and that does result in a AV that can be nearly 10% higher. And for your entertainment, there is not a restriction that the coinsurance be between 0% and 100%, so this bug can be used to get AV below 0% and above 100%.

I don't think 1 is really an issue since you can click the box "Do not allow copays to exceed service unit cost" and avoid negatives. The real question is why this is an even an option, since like you said, in practice no one is cutting a check to the insurer.

2 is even more complicated than it looks. If I understand what you're saying correctly, you think that a copay plan should enter the coinsurance as 0% instead of 100% otherwise the results will be misleading. But if you enter all $0 copays and make the coinsurance 0%, then the AV is around 82% instead of 100%. This nonsense occurs because the sum of the services is less than the total cost; and the calculator applies the coinsurance to the difference.

thing
12-04-2012, 05:37 PM
I don't think 1 is really an issue since you can click the box "Do not allow copays to exceed service unit cost" and avoid negatives. The real question is why this is an even an option, since like you said, in practice no one is cutting a check to the insurer.
I agree with this assessment. :D


2 is even more complicated than it looks. If I understand what you're saying correctly, you think that a copay plan should enter the coinsurance as 0% instead of 100% otherwise the results will be misleading. But if you enter all $0 copays and make the coinsurance 0%, then the AV is around 82% instead of 100%. This nonsense occurs because the sum of the services is less than the total cost; and the calculator applies the coinsurance to the difference.
I think the results are misleading no matter what you put in, currently. But the more I think about it, you're right, it's not obvious that a 0% coinsurance is a 'fix'.

afg5008
12-05-2012, 09:01 AM
So...has anyone coded this to batch rate multiple benefit designs? If so, care to share so that I can be saved the time to do this?

I'm probably missing something, but what would be the advantage of this? Let's say I have macro that takes all my company's plan designs and runs them through the calculator and spits out their AVs. To use the same macro, you'd need to put your plan designs in the same format as mine. And if you're going to go through that trouble, you might as well just put them directly into the AVC.

FormLetter
12-05-2012, 09:09 AM
Someone may be able to create a template that maps their plan design format into yours for input into the calculator via a batch process.

If they have hundreds or thousands of plan designs to calculate AVs for, it could easily be worth it.

KG004
12-06-2012, 10:41 AM
Does anyone know the VBA password by chance? The worksheet passwords I know, but any luck on the VBA yet?

thing
12-06-2012, 11:10 AM
Does anyone know the VBA password by chance? The worksheet passwords I know, but any luck on the VBA yet?

See post #15 of this thread.

FauxPas
12-06-2012, 11:25 AM
VBA errors aside, does anyone find the continuance tables themselves to be extremely troubling?

I understand the Cost per Enrollee (Max'd) column and agree with how it's used to value deductibles and MOOPs when all services are subject to the deductible.

The problems with their calcs arise when one starts covering services outside of the deductible. The service level columns are described in the methodology document this way:

"For each service type listed above, the columns of the continuance table display the average cost of spending on that service type that is attributed to cumulative enrollees in each range..."

I interpret "each range" to be the total allowed cost ranges (first column) in which members are bucketed. Then, I interpret, for example, the "ER" value for the "$500" row to be the average ER allowed cost for members with $500 or less in total costs.

This is the ER value that would get used in valuing a plan that covers ER outside the deductible (for instance, under a copay). But this seems totally wrong. What the ER value would need to be, to be correctly used by them in the VBA, is the average amount of ER allowed costs in the first $500 in total claims for all members. In practice, this is difficult to determine (one would have to sequence claims by member, figure out a straddle methodology, etc), which is why if I were valuing a plan with ER outside the deductible, I would calculate a whole separate continuance table with all costs except ER, and value my deductible that way, then add the net ER costs on later, and adjust my MOOP to reflect the ER copays if they count.

Am I crazy?

(I won't even touch on the fact that the service level values seem to be strictly increasing, which should not always be the case, based on their description and my interpretation.)

KG004
12-06-2012, 11:39 AM
See post #15 of this thread.

I have that, just wondering if anyone had the VBA password.

thing
12-06-2012, 12:22 PM
I am aware of various methods for removing a VBA password, but not aware of any methods for determining a VBA password. Which does not definitively answer your question.

thing
12-06-2012, 04:25 PM
Unofficial, and second hand, word: they will not be issuing a revised AV calculator prior to the close of the comment period.

TookTheScenicRoute
12-07-2012, 02:15 PM
I'm confused about the lesser of logic. Aside from how it's valued, because I can grasp the concept of valuing the lesser of two things, I'm confused about how it applies to a covered member benefits. If I buy a drug that costs $3 and my copay is $20, when would I ever pay $20, or am I misinterpreting? Who gets that $17 difference if it's not "lesser of"?

FormLetter
12-07-2012, 03:05 PM
You wouldn't ever pay the copay amount if the copay was greater than the drug price. Clicking that checkbox is you telling the calculator not to act as though you would.

TookTheScenicRoute
12-07-2012, 04:20 PM
My point exactly. If you click the button because that would not happen, then why is there a button that must be clicked to prevent valuing a nonexistent event? There has to be some instance of a benefit design where the button is unchecked, and I want to understand how that benefit design works. I'm hoping that someone can clear this up for me.

Reik483
12-07-2012, 05:11 PM
I'm gonna go with Occam's Razor on this one. If my two choices are

1. The person making this didn't really understand how insurance policies are designed.
2. There exist crazy insurance plans that make the member pay the full copay for 3 dollar drugs.

I'm going to go with the first one.

thing
12-07-2012, 06:20 PM
I'm gonna go with Occam's Razor on this one. If my two choices are

1. The person making this didn't really understand how insurance policies are designed.
2. There exist crazy insurance plans that make the member pay the full copay for 3 dollar drugs.

I'm going to go with the first one.

3. The model designer thinks a staff model HMO is still the predominant model.

accolade82
12-10-2012, 11:33 AM
Does anybody have a good sense as to what the regs mean concerning the flexability the actuary has to certify an AV for a plan that is outside the scope of the AV calculator?

Our current standard products have quite a few differences from the calculator (as I am sure everyone else does as well): Urgent Care Copay, Retail/Mail order Split, Generic Penalty, etc. I'd imagine that in aggregate these differences are material in calculating AV, even if each is immatierial on it's own.

Are we supposed to use the standard population data, or can we make adjustments based on our own data?

Also, anyone know how it handles the 2x family cap? All the data is at the member level and I am not sure if its considered. Its about a point difference in AV between assuming no family cap and assuming a 2x family cap.

DoctorNo
12-10-2012, 12:41 PM
Are we supposed to use the standard population data, or can we make adjustments based on our own data?

The actuarial adjustments cannot involve population differences.

DoctorNo
12-10-2012, 02:58 PM
Is there guidance about how to use the AV Calculator to price silver plan variations? I don't recall seeing anything (for instance, about which continuance table to use).

thing
12-10-2012, 03:29 PM
Does anybody have a good sense as to what the regs mean concerning the flexability the actuary has to certify an AV for a plan that is outside the scope of the AV calculator?

Our current standard products have quite a few differences from the calculator (as I am sure everyone else does as well): Urgent Care Copay, Retail/Mail order Split, Generic Penalty, etc. I'd imagine that in aggregate these differences are material in calculating AV, even if each is immatierial on it's own.

Are we supposed to use the standard population data, or can we make adjustments based on our own data?

My read is that you can use your own data to estimate how some of those variations map into what you input into the AV calculator, and that this is the best way of dealing with those. E.g. if you have Urgent Care copay of $50 and PCP copay of $25, you think both of those fall under the PCP line of the AV calculator, and your data is that 95% of visits are PCP visits, then it is appropriate to put a copay of 0.95*25+.05*50=26.25 on that line.

thing
12-10-2012, 03:34 PM
Is there guidance about how to use the AV Calculator to price silver plan variations? I don't recall seeing anything (for instance, about which continuance table to use).

VBA code implies use silver for 200% FPL and above, and platinum for below 200% FPL variations.


ElseIf Sheet1.CheckBoxes("csr_ind").Value = 1 And (numerator / denominator) >= 0.69 And (numerator / denominator) <= 0.71 Then
Sheet1.Range("B57").Value = "Silver"
Sheet1.Range("B55").Value = "CSR Level of 70% (250-400% FPL)"
ElseIf Sheet1.CheckBoxes("csr_ind").Value = 1 And (numerator / denominator) >= 0.72 And (numerator / denominator) <= 0.74 Then
Sheet1.Range("B57").Value = "Silver"
Sheet1.Range("B55").Value = "CSR Level of 73% (200-250% FPL)"
ElseIf Sheet1.CheckBoxes("csr_ind").Value = 1 And (numerator / denominator) >= 0.86 And (numerator / denominator) <= 0.88 Then
Sheet1.Range("B57").Value = "Platinum"
Sheet1.Range("B55").Value = "CSR Level of 87% (150-200% FPL)"
ElseIf Sheet1.CheckBoxes("csr_ind").Value = 1 And (numerator / denominator) >= 0.93 And (numerator / denominator) <= 0.95 Then
Sheet1.Range("B57").Value = "Platinum"
Sheet1.Range("B55").Value = "CSR Level of 94% (100-150% FPL)"
ElseIf Sheet1.CheckBoxes("csr_ind").Value = 1 Then
Sheet1.Range("B55").Value = "Error: Result is outside of +/- 1 percent de minimus variation HHS intends to propose for CSRs."
Exit Sub

DoctorNo
12-10-2012, 03:43 PM
Very interesting - thanks!

Do you have to tell the calculator that you're intending to price a variant? (I haven't poked around in the calculator in a bit).

EDIT: Never mind - I think your code snippet tells me that as well.

Mr F
12-11-2012, 11:44 AM
Has anyone played with the Primary Care Deductible/Coinsurance After a Set Number of Copays input? The intent is for the first PCP visit to be copay, and then deductible/coinsurance thereafter.

I have the deductible/coinsurance boxes unchecked and a copay entered in the Primary Care line, with the "Begin PCP Ded/Coins..." box checked and 1 copay down below. I would read this as the PCP line is being brought in for the first visit (copay only) and global ded/coins thereafter. This gives me an AV within the range I'm shooting for.

However, if I check the ded/coins boxed in the PCP line then the AV increases by more than 10%. That seems...odd. If the previous calculation was only bringing in copays for all PCP visits because I hadn't checked the deductible/coinsurance boxes, then I would think checking those boxes and subjecting PCP visits to deductible/coinsurance would lower the AV, not increase it 10%. (The ded/coins levels in this particular example are less rich than the somewhat low copay I have entered.)

Furthermore, if I take the same plan design and subject it to only copay without checking the "Begin PCP Ded/Coins..." box below then I get something 10-12% higher as well.

tl;dr version, with all other inputs being held constant:
PCP ded/coins begins after 1 copay, copay entered and ded/coins boxes not checked: x%
PCP ded/coins begins after 1 copay, copay entered and ded/coins boxes checked: over x+10%
PCP copay only, no PCP ded/coins after # of copays: over x+10%

I"m having trouble wrapping my head around what is causing this wild swing in AV. Anyone have any insights?

db20
12-11-2012, 12:18 PM
Has anyone played with the Primary Care Deductible/Coinsurance After a Set Number of Copays input? The intent is for the first PCP visit to be copay, and then deductible/coinsurance thereafter.

I have the deductible/coinsurance boxes unchecked and a copay entered in the Primary Care line, with the "Begin PCP Ded/Coins..." box checked and 1 copay down below. I would read this as the PCP line is being brought in for the first visit (copay only) and global ded/coins thereafter. This gives me an AV within the range I'm shooting for.

However, if I check the ded/coins boxed in the PCP line then the AV increases by more than 10%. That seems...odd. If the previous calculation was only bringing in copays for all PCP visits because I hadn't checked the deductible/coinsurance boxes, then I would think checking those boxes and subjecting PCP visits to deductible/coinsurance would lower the AV, not increase it 10%. (The ded/coins levels in this particular example are less rich than the somewhat low copay I have entered.)

Furthermore, if I take the same plan design and subject it to only copay without checking the "Begin PCP Ded/Coins..." box below then I get something 10-12% higher as well.

tl;dr version, with all other inputs being held constant:
PCP ded/coins begins after 1 copay, copay entered and ded/coins boxes not checked: x%
PCP ded/coins begins after 1 copay, copay entered and ded/coins boxes checked: over x+10%
PCP copay only, no PCP ded/coins after # of copays: over x+10%

I"m having trouble wrapping my head around what is causing this wild swing in AV. Anyone have any insights?

I'm not getting the same result for that scenario. The only difference is that you're just ticking that one box in row 51? I'm getting an increase from x to x+1.4%

Mr F
12-11-2012, 01:28 PM
I'm not getting the same result for that scenario. The only difference is that you're just ticking that one box in row 51? I'm getting an increase from x to x+1.4%

Scenario 1: Boxes in row 20 unchecked, copay in row 20, box in row 51 checked, 1 copay in row 52. Result: X

Scenario 2: Same as Scenario 1, boxes in row 20 checked. Result: x + 10%

Scenario 3: Same as scenario 1, box in row 51 unchecked with entry in row 52 empty. Result: x + 12%

For reference, the Bronze continuance table is being used. The results are coming from an original released locked version (so I haven't done anything else in the code to screw with it).

So the difference in Scenarios 1 and 3 is the update to rows 51-52. The difference in Scenarios 1 and 2 is just checking the deductible/coinsurance boxes in row 20.

jone2219
12-14-2012, 09:10 AM
i'm late to this game, but if you need to unlock the workbook and any individual sheets, pm me. im not guaranteeing anything though

FormLetter
12-14-2012, 09:40 AM
Do you have the password?

jone2219
12-14-2012, 09:44 AM
no....idk how to actually get it. do you?

FormLetter
12-14-2012, 10:28 AM
No I don't have it either.

jone2219
12-14-2012, 11:13 AM
Try: AABBABABAAGA


no joke....it worked for me

TwoStep
12-14-2012, 11:45 AM
Try: AABBABABAAGA


no joke....it worked for me

He (or she?) is serious. I did something similar in the past. There is a macro floating around online that does a brute force hack to find a password that works (all though that is a password, not the one they used to lock it). I'll try to find the macro.

You can also open the Excel file in OpenOffice.org's 'Calc' spreadsheet software, and it will unlock both the workbook and the VBA.

jone2219
12-14-2012, 12:17 PM
He (or she?) is serious. I did something similar in the past. There is a macro floating around online that does a brute force hack to find a password that works (all though that is a password, not the one they used to lock it). I'll try to find the macro.

You can also open the Excel file in OpenOffice.org's 'Calc' spreadsheet software, and it will unlock both the workbook and the VBA.

My guess is that the excel encryption has a many to one thing going on, so there are millions of potential actual passwords that could be hacked with the ABABA...password that my code produced.

Sub Crackwksht()

Dim v1 As Integer, u1 As Integer, w1 As Integer
Dim v2 As Integer, u2 As Integer, w2 As Integer
Dim v3 As Integer, u3 As Integer, w3 As Integer
Dim v4 As Integer, u4 As Integer, w4 As Integer
On Error Resume Next


For v1 = 65 To 66: For u1 = 65 To 66: For w1 = 65 To 66
For v2 = 65 To 66: For u2 = 65 To 66: For w2 = 65 To 66
For v3 = 65 To 66: For u3 = 65 To 66: For w3 = 65 To 66
For v4 = 65 To 66: For u4 = 65 To 66: For w4 = 32 To 126


ActiveSheet.Unprotect Chr(v1) & Chr(u1) & Chr(w1) & _
Chr(v2) & Chr(u2) & Chr(v3) & Chr(u3) & Chr(w3) & _
Chr(v4) & Chr(u4) & Chr(w4) & Chr(w2)


Next: Next: Next: Next: Next: Next
Next: Next: Next: Next: Next: Next



End Sub

Sub Crackwrkbk()


Dim v1 As Integer, u1 As Integer, w1 As Integer
Dim v2 As Integer, u2 As Integer, w2 As Integer
Dim v3 As Integer, u3 As Integer, w3 As Integer
Dim v4 As Integer, u4 As Integer, w4 As Integer
On Error Resume Next

For v1 = 65 To 66: For u1 = 65 To 66: For w1 = 65 To 66
For v2 = 65 To 66: For u2 = 65 To 66: For w2 = 65 To 66
For v3 = 65 To 66: For u3 = 65 To 66: For w3 = 65 To 66
For v4 = 65 To 66: For u4 = 65 To 66: For w4 = 32 To 126


ThisWorkbook.Unprotect Chr(v1) & Chr(u1) & Chr(w1) & _
Chr(v2) & Chr(u2) & Chr(v3) & Chr(u3) & Chr(w3) & _
Chr(v4) & Chr(u4) & Chr(w4) & Chr(w2)


Next: Next: Next: Next: Next: Next
Next: Next: Next: Next: Next: Next

End Sub

TwoStep
12-14-2012, 12:49 PM
My guess is that the excel encryption has a many to one thing going on, so there are millions of potential actual passwords that could be hacked with the ABABA...password that my code produced.

End Sub

Ya, I thought I remember reading that Excel's password protection hash is weak (so you'll see a lot of many to one relationships). I know VBA uses SHA1, which is a lot stronger.

jone2219
12-17-2012, 12:28 PM
how are people accounting for the family portion of their plans? i didnt find much clear guidance on this in anything i read, but i havent read all of it yet.

FauxPas
12-17-2012, 04:17 PM
how are people accounting for the family portion of their plans? i didnt find much clear guidance on this in anything i read, but i havent read all of it yet.

This was asked of John Bertko on Friday's Academy webinar on AV/EHB. My recollection is that he alluded to it being somewhat immaterial (I would disagree), but also noted that an adjustment could be made via actuarial certification. The AVC does not appear to consider per contract / per family in any way.

FormLetter
12-18-2012, 09:35 AM
It's only material if you consider it material. If your company wants to issue a plan for which metal level compliance can only be achieved by considering the family tier values for deductibles or other cost-sh provisions, then you go ahead and make adjustments with an actuarial certification and voila, a compliant plan design.

jone2219
12-19-2012, 03:04 PM
How are people getting around the run-time error if you have copays that aren't supposed to be applied to the ded or OOP? I have some 80% drug plans with certain copays, but there is no OOP.

lifishing007
12-19-2012, 03:12 PM
I believed the HHS is taking public comments on the AV calculator (of course not about the coding). Does anyone want clarification of any issues that has not been addressed before?

jone2219
12-19-2012, 03:13 PM
This was asked of John Bertko on Friday's Academy webinar on AV/EHB. My recollection is that he alluded to it being somewhat immaterial (I would disagree), but also noted that an adjustment could be made via actuarial certification. The AVC does not appear to consider per contract / per family in any way.

I was unable to sit in on the webinar...I have the handouts that came with it, but does anyone know if the Questions and Answers and recorded and available somewhere?

thing
12-19-2012, 03:30 PM
I believed the HHS is taking public comments on the AV calculator (of course not about the coding). Does anyone want clarification of any issues that has not been addressed before?

In last Friday's webinar, it was suggested that they are taking comments on the coding and appreciate those specific comments coming to the actuarialvalue@cms email address in advance of being sent in as official comments. That's how I heard it, anyway.

The announcement email indicates that the webinar was recorded and should be available "about a week" after the live event, from the same link they emailed you.

jone2219
12-19-2012, 04:13 PM
In last Friday's webinar, it was suggested that they are taking comments on the coding and appreciate those specific comments coming to the actuarialvalue@cms email address in advance of being sent in as official comments. That's how I heard it, anyway.

The announcement email indicates that the webinar was recorded and should be available "about a week" after the live event, from the same link they emailed you.

great, thanks

jone2219
12-19-2012, 04:20 PM
I just re-read all of these posts and I don't think I'm asking anything that was covered - sorry if I am...

A general comment about the interface:
What is row 42 (chronic maintenance drugs) used for? anything? nothing that I can tell at least...

Out of network costs do no count in the AV calculation. Does anyone know how indemnity plans are to be treated?
What is your understanding of this? I was under the impression that the OON benefits weren't under the same requirements as INN. Why is the OON OOPM being restricted to <= 6500? This wont fit well for many of my plan designs.


That's problematic. I assume the error would also mean it's looking at the (non-used) ded_pt and coins_pt checkboxes as well.

I couldn't find any of these un-used/hidden checkboxes. Where are they at?

Additionally, is there an easy way to get the list of the check-boxes and see where they are? The only way I know of is to right click on them individually and see what shows up in the named-range area.



What does everyone use as the annual contribution for the HSA/HRA plans? I'm new around here, but as far as I've heard this is something that is usually unknown to the insurer. But this has an effect on the final AV. I assume the same is true for the 1st/2nd tier utilization, but I would assume that those percentages are a bit easier to estimate based on your own company's experience on claims?

exponentialpi
12-19-2012, 04:53 PM
I was unable to sit in on the webinar...I have the handouts that came with it, but does anyone know if the Questions and Answers and recorded and available somewhere?

We inquired about the recordings as well due to mixed attendance and the response we received was that recordings will not be available as the government has decided not to allow the sessions to be replayed. For what its worth.

thing
12-19-2012, 07:20 PM
What is row 42 (chronic maintenance drugs) used for? anything? nothing that I can tell at least...

Doesn't seem to be used for anything, no.
I couldn't find any of these un-used/hidden checkboxes.[...] is there an easy way to get the list of the check-boxes and see where they are?


For Each my_cb In ActiveSheet.CheckBoxes
Debug.Print my_cb.Name
Next

lebron
12-22-2012, 01:09 PM
What is your understanding of this? I was under the impression that the OON benefits weren't under the same requirements as INN. Why is the OON OOPM being restricted to <= 6500? This wont fit well for many of my plan designs.

I don't think that is strictly OON. That is for another tier of providers you have contractual relationships with.

jone2219
12-23-2012, 01:51 AM
I don't think that is strictly OON. That is for another tier of providers you have contractual relationships with.

That's fine then, in the context of an alternative contracted provider network, but what about benefits when the member goes entirely out of network? Are providers still restricted to max cost sharing levels with the members when they get services oon?

swifthat
12-23-2012, 08:16 AM
That's fine then, in the context of an alternative contracted provider network, but what about benefits when the member goes entirely out of network? Are providers still restricted to max cost sharing levels with the members when they get services oon?

PPACA mentions that the following are not included in the word "cost-sharing" (in section 1301):

(B) EXCEPTIONS.—Such term does not include premiums, balance billing amounts for non-network providers, or spending for non-covered services

I'm not sure if any rules have been published that bring more clarity on this, but the exception sounds like OON is subject to the max cost sharing (with the exception of any balance billed amount). So maybe we'll be seeing ded/OOP equal for IN and OON in the future?

DoctorNo
12-26-2012, 10:43 AM
Today's the big day for comments to be due! :clap:

What are you commenting?

thing
12-26-2012, 11:27 AM
I got a couple of hours left - don't rush me... :-)

I'm thinking of putting in comments on two proposed rules today. I'm pretty upset with the single age curve idea.

meow_meow
01-08-2013, 11:49 AM
I'm very behind on the AV calculator.. but for those who have tested it out in great details, could you please share how you price for the inpatient professional services? all I see related to inpatient is "all inpatient hopital services" option... any thoughts on it would be greatly appreciated!

thing
01-09-2013, 11:37 AM
meow_meow: that's one of the issues currently - we have not been provided with an adequate description of what falls under each line of the AV calculator. Inpatient professional services are either on the line for inpatient hospital services, or split between that and one of the lines for primary care visits, specialist visits, and/or outpatient surgery physician/surgical services. Even if somebody knew for sure, having that information posted on an anonymous internet forum is of limited use. We need CMS to provide a lot more detail on how the mapping from claims to AV calculator categories was done.

Absent that, use your best judgment and document everything really well.

OnTheWayToFSA
01-09-2013, 05:28 PM
Does anyone have a link to the rules/regs that say the standard population (vs. company's own) needs to be used if you are going to make additional adjustments outside the AVC? I have heard that this is the case but can't find the supporting documentation.
And if this is true - don't we need more detailed info on this population (than what is in the AVC) to actually use it to make external adjustments?
Thanks!

DoctorNo
01-09-2013, 06:43 PM
Entering a plan design into the calculator where no services are subject to the deductible, one would expect that varying the deductible would not change the actuarial value of the plan (since nothing is subject to it). And yet, it does.

Does this mean:
(1) In the calculation, the listed categories do not form a partition of all health care costs (and there are categories not listed that are subject to the deductible)

Or

(2) This is an error.

For instance, one can create a plan where a $2,000 deductible only applies to SNF services, and the calculator calls it platinum. This plan would be essentially free to all members (and would cost a boatload, but that's another problem).

(Has this been discussed in this thread, or elsewhere, already?)

thing
01-09-2013, 09:46 PM
(3) All of the above.

There are costs that do not fall onto any of the lines of the calculator (about $260 per member per year off the top of my head, based on comparing the ultimate line of the continuance tables parts + 260 = total (roughly) (not at my work desk, etc.)) If we are being generous to the designer, that represents pediatric vision and dental costs, which had to be added in separately, and are an exception, etc. Except we can't explicitly put in cost sharing for them in the released version of the AV calculator.

But I think it's fairly clear that the way the calculator handles the deductible/oop/coinsurance is disconnected from reality as well, i.e. it's just bad at certain kinds of plan designs (copay only, anyone?)

[/rant]

DoctorNo
01-09-2013, 10:31 PM
That's what I was afraid of. Thanks!

cincinnatikid
01-10-2013, 03:17 PM
Does anyone have a link to the rules/regs that say the standard population (vs. company's own) needs to be used if you are going to make additional adjustments outside the AVC? I have heard that this is the case but can't find the supporting documentation.
And if this is true - don't we need more detailed info on this population (than what is in the AVC) to actually use it to make external adjustments?
Thanks!

You might be able to find a good lawyer to argue otherwise, but the intent based on everything I've read in the regs is that everything should use the 'standard population' as provided in the AV calculator, unless (starting in 2015) a state receives approval to use their own data.

http://www.gpo.gov/fdsys/pkg/FR-2012-11-26/pdf/2012-28362.pdf


AV will be calculated using the default standard population described in paragraph (f) of this section, unless a data set in a format specified by HHS that can support the use of the AV calculator as described in paragraph (a) of this section is submitted by a state and approved by HHS consistent with paragraph (d) of this section by a date specified by HHS.

OnTheWayToFSA
01-11-2013, 04:26 PM
You might be able to find a good lawyer to argue otherwise, but the intent based on everything I've read in the regs is that everything should use the 'standard population' as provided in the AV calculator, unless (starting in 2015) a state receives approval to use their own data.

http://www.gpo.gov/fdsys/pkg/FR-2012-11-26/pdf/2012-28362.pdf

Thanks, I was reading those sub-sections (156.135 e and f) as more pertaining to when you use the AV calculator (for non-exception plans) and since a specific reference was not made in section (b) Exception to the use of the AV calculator about which population to use, it seemed to possibly (?) imply you would only need an actuary to certify that "the plan design was fit appropriately in accordance with generally accepted actuarial principles and methodologies" in terms of finding a equivalent benefit to put in the model.

I can see now that could be read both ways though.

Sternjäger
01-14-2013, 12:15 PM
I've notice the deductible issue with the Rx too. If I deselect all of the subject to deductible boxes on the Rx one would think the Rx deductible wouldn't matter.

FormLetter
01-16-2013, 01:29 PM
I just read some notes from a call with an HHS Director and one of the industry organizations (ACHP or AHIP, not in front of me right now) where the industry group said there are major concerns with the AV calculator, and HHS responding saying they are getting good feedback on the calculator.

NotUrAvgXBar
01-16-2013, 04:14 PM
I just read some notes from a call with an HHS Director and one of the industry organizations (ACHP or AHIP, not in front of me right now) where the industry group said there are major concerns with the AV calculator, and HHS responding saying they are getting good feedback on the calculator.

I think that "good feedback" means very detailed descriptions of all the things that are wrong with the tool.

grivera
02-05-2013, 05:42 PM
Does anyone know why I keep getting an error in the AV Calc whenever I enter a different coinsurance or copay for Specialist Visit and Rehab/Phys Therapy?

grivera
02-05-2013, 05:52 PM
Does anyone know why when I enter a different coinsurance/copay for Specialist visit and Rehab/Physical Therapy I get an error? It looks like it only happens on the Specialist Visit coinsurance/copay side when Integrated Med & RX checkbox is NOT selected, and on the Rehab side when the boc IS selected???

GhostMac
02-07-2013, 06:12 PM
i brought up these errors awhile ago. no response from the forumn.

GhostMac
02-07-2013, 06:13 PM
error also comes up when you uncheck Rx coinsurance for non integrated in a 2 tier calculation.

health.actuary
02-10-2013, 05:27 PM
Please PM me serious issues you see with the calculator.

thing
02-10-2013, 05:41 PM
Please PM me serious issues you see with the calculator.

Are you in a position to do something about them?

algebrat
02-10-2013, 05:46 PM
I just saw this thread. Would looking at this program be a good way to learn abou the job ahead of time?

daaaaabears
02-11-2013, 02:04 AM
Probably a good way of looking at the future of health actuaries (in that the govt will be determining a plans av by their own shitty algorithms), not good if you have no experience and don't know why it's coded as it is.

Although I guess it's a better-than-nothing way to learn about actuarial value.

Maximus
02-11-2013, 09:22 AM
After running all of our plans through the AV calculator and making a comparison of the AV as calculated by the the HHS calculator to the actual AV (Paid / (Paid + cost share amounts paid by the member)), it seems the HHS AV calculator actually does a pretty good job. Although, our company performed this exercise only on our New York business, so I can not speak for how this comparison would work for plans issued outside of New York.

thing
02-11-2013, 11:53 AM
After running all of our plans through the AV calculator and making a comparison of the AV as calculated by the the HHS calculator to the actual AV (Paid / (Paid + cost share amounts paid by the member)), it seems the HHS AV calculator actually does a pretty good job.

:iatp:

It's not 100% broken, no. Doing the same exercise, I haven't gotten any variances of more than 10%, and usually within 3%.

New York Actuary
02-11-2013, 09:20 PM
We have found the calculator to generally produce accurate AVs as well.

DoctorNo
02-11-2013, 09:24 PM
It seems to do best for the higher (platinum and gold) value plans.

Having said that, if you look at it for its real purpose - categorizing plans in a consistent fashion - it does a great job. Even if it misses the mark on all bronze plans, it's missing the mark consistently, and they'll all get put together.

thing
02-13-2013, 05:30 PM
regtap.info just sent email for two webinars on AV/EHB, for 2/20 and 3/6. From this, I optimistically conclude that the 'fixed' AV calculator will be available prior to 2/20.

Don't disabuse me, I'm in my happy place. Ok, my slightly less stressed out place. Ok, my differently stressed out place.

GhostMac
02-14-2013, 09:37 AM
Nevermind. I registered and am waitlisted.

FormLetter
02-14-2013, 01:11 PM
:iatp:

It's not 100% broken, no. Doing the same exercise, I haven't gotten any variances of more than 10%, and usually within 3%.

When you say 10%, I'm assuming you mean...
Plan A has a "true" AV of 80%, and the HHS AV calculator produces a resulting between 70% and 90%, right?
...NOT that you mean...
Plan A has a "true" AV of 80%, and the HHS AV calculator produces a resulting between 72%% and 88%, right?

Just curious.

thing
02-14-2013, 02:37 PM
When you say 10%, I'm assuming you mean...
Plan A has a "true" AV of 80%, and the HHS AV calculator produces a resulting between 70% and 90%, right?

Yes, this.

New York Actuary
02-14-2013, 05:27 PM
You would also need to have a sufficiently high number of claims in a given plan in order to appropriately make an assessment.

FormLetter
02-14-2013, 09:18 PM
You would also need to have a sufficiently high number of claims in a given plan in order to appropriately make an assessment.

Disagree.

New York Actuary
02-14-2013, 11:17 PM
Disagree.

My point is that in order to make a valid comparison between the HHS AV calculator value and the actual AV for a specific plan (amount paid by the insurer / (amount paid by the insurer + amount paid by the enrollee), it is not going to work if you only have a handful of claims in that specific plan. Surely you don't disagree with that, right?

FormLetter
02-15-2013, 07:40 AM
The HHS calculator is giving you an EV(AV) for the year 2014. You can't compare actual to expected until a few months into 2015 at the earliest. But you can compare actual to expected whether you have 12 member months, 1200 or 12000 member months.

(At first I thought you were talking about comparing the HHS calculator's EV(AV) with your own model's EV(AV), which I submit you can do regardless of how many claims occur in any particular plan design.)

afg5008
02-15-2013, 11:36 AM
It seems to do best for the higher (platinum and gold) value plans.

Having said that, if you look at it for its real purpose - categorizing plans in a consistent fashion - it does a great job. Even if it misses the mark on all bronze plans, it's missing the mark consistently, and they'll all get put together.

That hasn’t been my experience. There are too many important plan design features that the model can’t handle, and there isn’t enough guidance on how to consistently model plans. For instance, I have a plan that is all copays. I can’t even enter copays for the OP surgery, which is worth about a quarter of the AV value!

An equivalent coinsurance amount can be entered, but different actuaries are likely to come up with very different estimates for what that equivalent amount should be, especially since it isn’t even clear what claims are rolled up into outpatient surgery.

Then there’s the issue with the deductible and coinsurance amounts that need to be entered at the top of the sheet. Since it’s a copay only plan, those should have zero effect on the AV of the plan. But instead, I am seeing the AV change between 89% and 78% depending on what coinsurance values are entered. As it stands now, competitors could offer completely identical plans and reasonably calculate AVs that differ by an entire metal tier.

Considering that the whole point, as you mentioned, was to value plans consistently and not have it vary by the different assumptions that might be used by different companies/actuaries, I would say that this tool fails pretty hard.

FormLetter
02-15-2013, 02:25 PM
Then there is this question: Let's say you have three different plan designs, A, B, and C, and they all register AV=70% according to the HHS calculator.

However, when you use your proprietary model, you get:
AV(A) = 68%, AV(B) = 70%, and AV(C) = 72%

Which plan design would you prefer to put out in the market, and why?

GhostMac
02-18-2013, 10:44 AM
Question on X-rays and Diagnostic Imaging: For a 100%/Ded waived plan, why does the AV calculator produce very different results when the benefit's D&C toggles are left unchecked vs if D&C are checked?

Shouldn't no ded/coins be the same as $0 ded/100%coins?

thing
02-18-2013, 11:46 AM
The final sweep, looking at claims between deductible/MOOP, would be different depending on whether you toggle the checkboxes. Not appropriately, I might add, but that's probably what's causing the difference.

thing
02-18-2013, 11:49 AM
Then there is this question: Let's say you have three different plan designs, A, B, and C, and they all register AV=70% according to the HHS calculator.

However, when you use your proprietary model, you get:
AV(A) = 68%, AV(B) = 70%, and AV(C) = 72%

Which plan design would you prefer to put out in the market, and why?
Do I care, as long as all three are priced correctly?

GhostMac
02-18-2013, 12:55 PM
The final sweep, looking at claims between deductible/MOOP, would be different depending on whether you toggle the checkboxes. Not appropriately, I might add, but that's probably what's causing the difference.

Do you know which method would be correct for a copay plan?

D&C unchecked or D&C checked?

Thanks for the help!

FormLetter
02-18-2013, 01:10 PM
Do I care, as long as all three are priced correctly?

The phrase "priced correctly" can mean just about anything.

I think you would care, absolutely.

thing
02-18-2013, 01:55 PM
Why I might prefer the 68% AV plan - higher member cost share than is assumed in the risk adjustment mechanism, probably a lower premium and so might sell better.

Why I might prefer the 70% AV plan - tighter correlation with CMS assumptions means I'm less exposed to unintended consequences, and am more confident in my pricing versus other metal levels.

Why I might prefer the 72% AV plan - narrower gap between standard and CSR version of plans means less chance of underpricing the whole package of silver plans, higher likely premium more reflective of selection risk between Bronze and Silver

So, no, I don't think I care as long as they're "priced correctly", whatever that means...

thing
02-18-2013, 07:06 PM
regtap.info just sent email for two webinars on AV/EHB, for 2/20 and 3/6. From this, I optimistically conclude that the 'fixed' AV calculator will be available prior to 2/20.

Don't disabuse me, I'm in my happy place. Ok, my slightly less stressed out place. Ok, my differently stressed out place.
And then, today:

Thank you for registering for the EHB/AV Q&A User Group Session Series. Please be advised the February 20, 2013 session will be rescheduled. You will remain confirmed for this series and updated dates will be provided soon.

FormLetter
02-18-2013, 07:13 PM
The best part is that when the QHP certification applications go in, you have to have network, plan design, and rating information included, and that is by end of April. But final rules aren't out yet. And state insurance departments haven't responded to those as of yet unreleased final rules yet. And insurance companies haven't had all of that information to work with. It's borderline funny.

thing
02-18-2013, 07:17 PM
States have, what, 30 days after final rules to request variances? So rulings on those take how long? It is pretty much impossible for us to know the final rules prior to submission.

Yep. Hilarious.

health.actuary
02-19-2013, 10:15 AM
If people have questions/inquiries about the AVC please PM me. You might get answers.

New York Actuary
02-19-2013, 11:16 AM
If people have questions/inquiries about the AVC please PM me. You might get answers.

Are you with HHS? Are you in a position to do something about these issues? If so, why are you asking that these issues be taken off-line?

health.actuary
02-19-2013, 12:25 PM
Are you with HHS? Are you in a position to do something about these issues? If so, why are you asking that these issues be taken off-line?

If I wanted to get into my employment situation I would. I'm in a posistion where I might be able to help some people and getting feedback on issues/errors would be helpful to me. Some questions I can answer and some questions I can't.

To the people who PMed me:

The more detailed the question the better. The fewer acronyms the better.

thing
02-19-2013, 06:44 PM
If anybody from HHS wants to leak information on the pending 'fixed' AV calculator, even just "probably by this date", feel free to create a completely new account, PM me the information, and I will post it in this thread as being from an anonymous source.

If anybody not from HHS has questions about the original AV calculator, I would guess you are as likely or more likely to get an answer by posting it in this thread than by PMing anybody also not from HHS.

:shrug:

New York Actuary
02-20-2013, 08:53 AM
If anybody from HHS wants to leak information on the pending 'fixed' AV calculator, even just "probably by this date", feel free to create a completely new account, PM me the information, and I will post it in this thread as being from an anonymous source.

If anybody not from HHS has questions about the original AV calculator, I would guess you are as likely or more likely to get an answer by posting it in this thread than by PMing anybody also not from HHS.

:shrug:

:iatp:

Zee
02-20-2013, 10:28 AM
If anybody from HHS wants to leak information on the pending 'fixed' AV calculator, even just "probably by this date", feel free to create a completely new account, PM me the information, and I will post it in this thread as being from an anonymous source.

If anybody not from HHS has questions about the original AV calculator, I would guess you are as likely or more likely to get an answer by posting it in this thread than by PMing anybody also not from HHS.

:shrug:

:iatp: as well. Free flow of information on this topic, please!

cincinnatikid
02-20-2013, 10:39 AM
The rules have been at the White House pending final approval since the beginning of last week. I've been told that they have now been approved and are just awaiting publication (maybe Friday?). This doesn't necessarily mean the final calculator will be released simultaneously, but everything I've heard leads me to believe its release is imminent.

FormLetter
02-20-2013, 10:55 AM
The rules go to the White House? Where is Sebelius's office?

cincinnatikid
02-20-2013, 12:04 PM
http://www.ofr.gov/OFRUpload/OFRData/2013-04084_PI.pdf

New York Actuary
02-20-2013, 12:20 PM
http://www.ofr.gov/OFRUpload/OFRData/2013-04084_PI.pdf

Thanks Cincinnati Kid!

There does not seem to be an actual Excel AV calculator, just a bunch of XML files. Am I just missing it?

cincinnatikid
02-20-2013, 12:29 PM
Thanks Cincinnati Kid!

There does not seem to be an actual Excel AV calculator, just a bunch of XML files. Am I just missing it?

http://cciio.cms.gov/resources/files/av-calculator-final-2-20-2013.zip.zip

cincinnatikid
02-20-2013, 12:35 PM
http://cciio.cms.gov/resources/files/av-calculator-final-2-20-2013.zip.zip

Yeah, seems to be all .xml files in there...not sure what to do with that.

cincinnatikid
02-20-2013, 12:36 PM
Yeah, seems to be all .xml files in there...not sure what to do with that.

http://cciio.cms.gov/resources/files/av-calculator-final-2-20-2013.xlsm

thing
02-20-2013, 12:44 PM
I'm attempting to crack the new one now... hang on...

thing
02-20-2013, 12:44 PM
I'll be damned - they didn't bother locking! Thanks, HHS!!!!

DoctorNo
02-20-2013, 01:00 PM
Interesting stuff in here so far...

brooksyj
02-20-2013, 01:00 PM
I'll be damned - they didn't bother locking! Thanks, HHS!!!!

So where's the spreadsheet? Anyone? Any news?

exponentialpi
02-20-2013, 01:03 PM
So where's the spreadsheet? Anyone? Any news?


CCIIO Link (http://cciio.cms.gov/resources/regulations/index.html#pm)

thing
02-20-2013, 01:16 PM
So where's the spreadsheet? Anyone? Any news?
cincinattikid linked to it just above. I'm going through the issues I told them about, so far looks like they actually fixed things...

brooksyj
02-20-2013, 01:18 PM
Thanks, exponentialpi. That worked better. But this is new? It looks just like the old one. I was hoping for at least a line for DME and other EHBs.

FormLetter
02-20-2013, 01:20 PM
You 'hoped' for it, but you didn't 'expect' that, did you?

thing
02-20-2013, 01:24 PM
Looks like the lesser_of checkbox has been hidden, but left unchecked, and it is still referenced in the code. This could be a major problem.

thing
02-20-2013, 01:27 PM
Not obviously. I can't force a negative AV using the same trick as previously.

I'm a little too pumped up about having this to play with. Something is seriously wrong with me...

exponentialpi
02-20-2013, 01:27 PM
Thanks, exponentialpi. That worked better. But this is new? It looks just like the old one. I was hoping for at least a line for DME and other EHBs.

The section title is today's date and my file download has today's date in the file name. I also believe this is the first release of the MV calculator.

lifishing007
02-20-2013, 01:32 PM
I'm a little too pumped up about having this to play with. Something is seriously wrong with me...

It is fun. :)

brooksyj
02-20-2013, 01:42 PM
The section title is today's date and my file download has today's date in the file name. I also believe this is the first release of the MV calculator.

Okay, the values come up a little different with the same inputs. Why don't they tell us what they changed? Why do I bother to ask?

FormLetter
02-20-2013, 01:47 PM
Not obviously. I can't force a negative AV using the same trick as previously.

I'm a little too pumped up about having this to play with. Something is seriously wrong with me...

I'm curious what kinds of things you're trying to figure out by playing with it.

thing
02-20-2013, 01:47 PM
If numerator / denominator > 1 Or numerator / denominator < 0 Then
Sheet1.Range("B55").Value = "Error: Plan as specified returns an invalid AV."


:lol:

Cheaters!

FormLetter
02-20-2013, 01:47 PM
So if I make a plan with $0 medical deductible, $0 drug deductible, 100% carrier insurer on both medical and drug, and a $1000 OOP Max, the AV is 91.1%. But there's no cost-sharing!

FormLetter
02-20-2013, 01:48 PM
:lol:

Cheaters!

Yeah, that message takes away all the fun.

The logic goes like this:

IF(calculator screws up massively, "invalid AV bye bye", AV)

lifishing007
02-20-2013, 01:50 PM
Does anyone know how to adjust for three-tiered products?

FormLetter
02-20-2013, 01:54 PM
Does anyone know how to adjust for three-tiered products?

There are a few ways you can try.

Enter tier 1 as tier 1, tier 2 as tier 2, and get an AV, call it X.
Then enter tier 1 as tier 1, tier 3 as tier 2, and get an AV, call it Y.
Then use some assumption to weight X and Y based on expected distribution of utilization between tiers 2 and 3.

Alternatively, ignore tier 3. That one is probably a little harder to justify.

Alternatively, enter tier 1 as tier 1, and a weight average of the cost-sharing provisions for tiers 2 and 3 and tier 2. This could get messy.

FormLetter
02-20-2013, 01:58 PM
So if I make a plan with $0 medical deductible, $0 drug deductible, 100% carrier insurer on both medical and drug, and a $1000 OOP Max, the AV is 91.1%. But there's no cost-sharing!

Ok, I just made a gold plan with no cost-sharing:

$0 medical deductible
$0 drug deductible
100% carrier coinsurance on medical
100% carrier coinsurance on drug
$6500 OOPM

AV = 81.0%

thing
02-20-2013, 01:59 PM
I'm curious what kinds of things you're trying to figure out by playing with it.

Couple of things...

1. I wrote a lot of code to automate running our current plan designs through the prior version of the calculator. It only took a minor tweak of that to get inputs to work for the new version as well; although I'm now having to turn off a few extra 'Exit Sub' statements to get AVs of our current block.

2. Last time around, I finally convinced myself of a 'right' combination of "lesser_of" and coinsurance on straight copay plans to get an AV that sort of agreed with my own pricing model. They presumably made big changes to how the coinsurance works on straight copay plans, as that issue was raised to them, and so need to verify that still matches what my own pricing model shows.

3. ???

4. Profit!

lifishing007
02-20-2013, 01:59 PM
There are a few ways you can try.

Enter tier 1 as tier 1, tier 2 as tier 2, and get an AV, call it X.
Then enter tier 1 as tier 1, tier 3 as tier 2, and get an AV, call it Y.
Then use some assumption to weight X and Y based on expected distribution of utilization between tiers 2 and 3.

Alternatively, ignore tier 3. That one is probably a little harder to justify.

Alternatively, enter tier 1 as tier 1, and a weight average of the cost-sharing provisions for tiers 2 and 3 and tier 2. This could get messy.

Knowing the calculator, what if a company enter it one way, come out Gold; the regulator enter it another one, come out Silver? Is there some guidance on the degree of freedom?

FormLetter
02-20-2013, 02:01 PM
Knowing the calculator, what if a company enter it one way, come out Gold; the regulator enter it another one, come out Silver? Is there some guidance on the degree of freedom?

That is why you need to be able to justify why you did something that the calculator doesn't handle in a straightforward manner.

thing
02-20-2013, 02:05 PM
Changes seen from last time so far between -4% and +7%... about 5% of plans repriced so far.

lifishing007
02-20-2013, 02:09 PM
That is why you need to be able to justify why you did something that the calculator doesn't handle in a straightforward manner.

What kind of guidance, if you are expecting any, from the state division of insurance?

FormLetter
02-20-2013, 02:11 PM
What kind of guidance, if you are expecting any, from the state division of insurance?

From the ones I deal with? Or the ones you deal with? Because I don't know which ones you deal with.

Most of mine will be pretty minimal, except for one that will always pick an extreme and make you prove why every difference between what you did and that extreme is correct. I suggested once that for that state, we submit a rate filing that said "decrease all rates by 40%", that way when they argue the extreme of raising 30%, we can respond "upon further consideration, Company BLARG agrees to the suggested rate increase".

FormLetter
02-20-2013, 02:12 PM
Changes seen from last time so far between -4% and +7%... about 5% of plans repriced so far.

Are those changes in the category of "ok, now THAT result looks more reasonable than the old one", or are they more like "I see they put extra effort into making things more wrong!" ??

lifishing007
02-20-2013, 02:13 PM
From the ones I deal with? Or the ones you deal with? Because I don't know which ones you deal with.

Most of mine will be pretty minimal, except for one that will always pick an extreme and make you prove why every difference between what you did and that extreme is correct. I suggested once that for that state, we submit a rate filing that said "decrease all rates by 40%", that way when they argue the extreme of raising 30%, we can respond "upon further consideration, Company BLARG agrees to the suggested rate increase".

so far, I am on the regulator's side (not HHS)

FormLetter
02-20-2013, 02:15 PM
so far, I am on the regulator's side (not HHS)

In that case, I recommend your "guidance" be the following:

Regulator: How did you get this AV, Mr. Insurance Man?
Mr IM: I used the AV calculator and made a tweak. Here is the Excel file.
Regulator: Thanks.

afg5008
02-20-2013, 02:19 PM
Ok, I just made a gold plan with no cost-sharing:

$0 medical deductible
$0 drug deductible
100% carrier coinsurance on medical
100% carrier coinsurance on drug
$6500 OOPM

AV = 81.0%

Replace 100% with 80% coinsurance and the AV goes up to 84.6%. This is going to be a nightmare.

lifishing007
02-20-2013, 02:19 PM
In that case, I recommend your "guidance" be the following:

Regulator: How did you get this AV, Mr. Insurance Man?
Mr IM: I used the AV calculator and made a tweak. Here is the Excel file.
Regulator: Thanks.

lol, I will let my boss know.

What kind of guidance, if you are expecting any, from the state division of insurance?

I ask this question to avoid over burdensome guidance. I really appreciate if anyone could provide more comments.

lifishing007
02-20-2013, 02:34 PM
Replace 100% with 80% coinsurance and the AV goes up to 84.6%. This is going to be a nightmare.

I second that.

Bamafan
02-20-2013, 03:03 PM
This code may be the culprit:

eff_coins = (eff_coins + 297.06) / (Worksheets(Sheetstr).Cells(88, "C").Value)
If coins < 1 And oop > deduct Then
oop_max = (oop - deduct) / (1 - coins) + deduct
ElseIf deduct >= oop Then
oop_max = deduct
Else
oop_max = (oop - deduct) / (1 - eff_coins) + deduct
coins = eff_coins
End If

From the VBA underlying the AV calculator; Module 1, Function CombinedMacro.

FormLetter
02-20-2013, 03:09 PM
So what happens now if there are problems/glitches with this version of the AV calculator, but the final rule has been published?