President Obama's "7M on Medicaid Thanks to the ACA" Now True!! (UPDATE x2)

I don't usually write this all out, but given the interest in the site these days, I figured I would give a play-by-play of the process of breaking down the CMS reports by state, using the February CMS report just released an hour or so ago.

First, skip to Pages 12-16, where you can see the grand total for February (which, unlike the HHS reports, covers February 1-28 itself, none of this "Saturday closest to the end of the month" business):

  • Expansion States Subtotal: 1,303,676
  • Non-Expansion States Subtotal: 945,444
  • Total: 2,249,120

Hmmm...that looks like a lot less than 3 million to me. The main reason for the missing 750K is because some of the new enrollments ran through the ACA exchanges (HC.gov/etc) while others went through traditional means (walking in the door of a state Medicaid agency office, etc). Furthermore, some of the states overlap--they're reported on both reports, which means I have to be careful not to double-count:

The Affordable Care Act created a “no wrong door” policy, which means that individuals can apply for health coverage through the Marketplace or the Medicaid or CHIP agency (if it is a separate agency) in their state. Regardless of which “door” they choose, individuals can get eligibility determinations for all types of health coverage, including financial assistance to help pay for coverage, and have their accounts routed to the program for which they are eligible. This means that for a full picture of Medicaid and CHIP activity, the numbers in this report—which come from the state level—need to be understood in concert with the numbers previously reported on Health Insurance Marketplace enrollment.1 

...in SBM states, the data include Medicaid and CHIP applications and determinations that were reported in the Health Insurance Marketplace: Enrollment Report (in other words, these data are reported in both reports). For Federally-Facilitated Marketplace (FFM) states, the applications included in this report do not include applications made to the FFM which are reported in the Health Insurance Marketplace: Enrollment Report.  

This is to say that there's around 750K already listed on the spreadsheet under the Feb. HHS column, although this has been further lopped down to around 226K, since some states have supplanted it with more recent data in a later column. Since 750K of the 3M have already been accounted for (either included on the spreadsheet or replaced with newer data from the exchanges), that leaves about 2.25M from the CMS report to deal with.

We're just getting started, however.

Next up: Renewals (aka Redeterminations). Most of the data in the CMS reports only includes those newly enrolled in Medicaid/CHIP, but some states mix the two together. For those states, I can't use any of the data, since I have no way of knowing what the ratio of renewals to newly added is. Further cluttering things up is that some states include renewals in the number of applications, but not in the number of determinations (I don't use the application data at all, but it did confuse me at first):

13 The following states have included renewals in their February 2014 application data: Nevada, New Mexico, Ohio, and Pennsylvania.
14 These states have included renewals in their February 2014 determination data: Alaska, District of Columbia, Iowa, Maryland, Michigan, Nevada, North Dakota, New Mexico, Rhode Island, South Dakota, Texas, Utah, and Virginia. Some of these renewals are conducted without the individual submitting a new application, using information already in the system that the individual is asked to verify. 

So, I have to eliminate the data from 13 states completely. How many people is this? Well, technically Rhode Island only includes renewals for Medicaid, not CHIP, but that's only a couple hundred anyway and RI has already been replaced by more recent data (as of this morning) anyway, so I can eliminate all of it. This means I have to subtract:

  • AK: 3,993
  • DC: 5,894
  • IA:  26,352
  • MD: 43,777
  • MI: 52,388
  • NV: 14,654
  • ND: 0
  • NM: 25,309
  • RI: 10,508
  • SD: 1,200
  • TX: 337,680
  • UT: 36,067
  • VA: 29,868

...or a total of 587,690. That's right: I'm subtracting over 580,000 people from the Spreadsheet and Graph purely because some of them are renewals of existing accounts, not newly added...but I have no idea how many, so I'm tossing them all out.

If anyone wants to accuse me of overestimating the percentages of "woodworkers" or "strict expansion" enrollees from the other states, just point this out to them. I have a 580K cushion to work with here.

OK, so that leaves us with  2,249,120 - 587,690 = 1,661,430 remaining. What's next?

Well, some of the states have had both their exchange AND non-exchange enrollments supplanted by more recent data from the exchanges themselves. These include:

  • Hawaii: 6,089
  • Washington: 0

And then there's California. If you look on Page 13 of the report, you'll see a whopping thirteen different footnotes. Not all of them are relevant here, but several are...I think. The one which leaps out to me is "Includes 145,661 individuals eligible via targeted enrollment strategy".

Now, CoveredCA's press release the other day gave the total Medicaid/CHIP/Medi-Cal categories as around 1.5 million...via the state Exchange, but possibly including other means as well. In this case, there's a whopping 483,161 to deal with, and I'm not sure how much overlap there is with the other figures I have. I know that the 650K LIHP transfers aren't included (those were done back in January, and that's a larger number than the 482K anyway). I'm not sure about some of the other totals, but I'm pretty sure that I have to subtract at least the 145,661 noted in that footnote, since it's almost certainly being double-counted.

  • Therefore, subtract another 145,661 from California's 483,161, leaving 337,500 for CA

Hopefully the other CA Medicaid numbers will be cleared up when the March HHS report is released in the next week or so.

So, this leaves us with: 1,661,430 - 6,089 - 145,661 = 1,509,680

But wait, there's more!!

You see, in 5 states (Alaska, Connecticut, Nevada, North Carolina and Oregon), they don't report individuals...they report households!

That means I have to apply my (fairly conservative) 1.8x multiplier factor to include spouses, kids, etc. This means nothing for Alaska or New Mexico since I had to eliminate them anyway, but it increases the other 3 states somewhat:

  • Connecticut: 11,614 x 1.8 = 20,905
  • North Carolina: 57,768 x 1.8 = 103,982
  • Oregon: Wait...what's this? "Includes 6,908 individuals determined eligible via Targeted Enrollment Strategy." Ah, right; Oregon's "Fast Track" program is already listed, so I have to subtract 6,908 from 38,884 first, for 31,976. Then I multiply that by 1.8x to get 57,556.

There's something like 50 more footnotes of one sort or another, regarding everything from "MAGI systems" to "non-title XIX programs including AIDS Drug Assistance", but most of the others aren't relevant here.

So, after subtracting the 6,089 from Oregon and then increasing CT, NC and OR from 101,358 to 182,443 to account for the "households" factor (an increase of 81,085), we're finally at our new total: 1,509,680 - 6,908 plus 81,085 = 1,583,557

Whew!

However, there's one more step left:

This only tells us how many new enrollees were added in February. It doesn't tell us how many of those were a) "strict expansion", b) "woodworkers" or c) "baseline churn".

Here's where we get into my "best guess" estimates.

For that, it depends on the state. Some states, such as Washington and Rhode Island, have done this for me; they've issued press releases which specifically state how many of the total new Medicaid enrollees are due to the actual Expansion provisions in the ACA.

For the other states, it gets trickier.

"Strict Expansion" is very easy for the non-expansion states...I simply enter none. Those cells are blanked out completely.

For the expansion states, which don't specifically tell me the number of "strict expansion" enrollees, I'm using this general rule of thumb:

  • 50% Strict Expansion
  • 20% Woodworkers
  • 30% Baseline Churn (not included at all)

For the non-expansion states, I'm using the following rule of thumb:

  • 15% Woodworkers
  • 85% Baseline Churn (not included at all)

Based on this, the totals for each category due to the February CMS report alone have gone from:

  • Strict Expansion
    • Was: 4,747,140
    • NOW: 5,023,440
  • Woodworkers
    • Was: 1,812,781
    • NOW: 2,012,548
  • Combined
    • Was: 6,559,921
    • NOW: 7,035,988

...for a total increase of 476,067.

Now, I'm sure I've missed a few things. I probably still have some double-counting here and there (I'm especially confused about that big 350K+ California number). However, as noted above, I've also deliberately deleted over a half a million potential enrollees due to the "renewals mixed in" factor...and that's from February alone. Similar numbers were purged from the total in the earlier months as well.

Therefore, I'm quite confident that any data which I have included which is double-counted or overestimated is more than cancelled out by the enrollees whcich I've had to delete entirely.

That's why I'm pretty confident that I have this pretty close.

UPDATE: Title changed to something less boring and more snarky. I just realized that this means President Obama's DGA speech gaffe from February 20 (he claimed at the time there had been "7 million Americans who have access to health care for the first time because of Medicaid expansion."), which got him into some hot water with several fact-checkers, is now true!!

OK, so he was about 6 weeks early, but still...

UPDATE x2: Several people have asked about the "3 Million More" figure being touted by Sec. Sebelius and the CMS. I'm assuming that this is the net increase in enrollments vs. last year...after subtracting people who have since left Medicaid due to having died, moved, gotten jobs or otherwise no longer need Medicaid for one reason or another.

The problem is, there's no way in hell I can track all of that, certainly not on the graph--it would be an impossible task, especially for a single person. I'm already overwhelmed by trying to track the increase in enrollments daily, much less the decrease on a daily, weekly or even monthly basis.

The best I can do is to keep the tally going until the next Open Enrollment period...at which point I will (assuming I'm still doing this) reset everything back to zero and start over.