Medicare Part D: How To Lie With Statistics

by hilzoy

It's been a while since I last posted about Medicare Part D. Today's topic: enrollment numbers. Kate Steadman at TPMCafe's Drug Bill Debacle blog pointed me to a WSJ article (sorry, subscription) that claims that the administration has been cooking its figures:

"Despite the headline on a government press release -- "30 million Medicare beneficiaries now receiving prescription drug coverage" -- a smaller number is actually enrolled in the new program, and some of that group had coverage before. As of April 18, 19.7 million beneficiaries are getting drug insurance from Medicare. Of that group, 5.8 million already had coverage from Medicaid, the state-federal program for the poor. An additional 6.8 million people are getting drug coverage from former employers; the coverage is partially subsidized by Medicare. That means a total of 26.5 million people now are benefiting from the Medicare drug program.

To get to 30 million, government officials also counted 3.5 million people who have drug coverage from the military's TRICARE program or federal-employee benefits, but aren't signed up for the Medicare benefit. Christina Pearson, a spokeswoman for the Department of Health and Human Services, says the Medicare beneficiaries, regardless of source of coverage, "were able to make the choice that works best for them" because of the new drug-benefit program."

Here's the press release. In it, HHS Secretary Mike Leavitt is quoted as saying:

"With a month to go, we've passed our projections of 28 to 30 million enrollees in the first year, and we are intensifying our local outreach efforts to get more seniors signed up before the May 15th deadline," HHS Secretary Mike Leavitt said. "These strong enrollment numbers are a tribute to everyone in the national network of caring -- all of our partners, community leaders, the State Health Insurance Programs (SHIPS), and family members, who have provided counseling and assistance to the millions of beneficiaries who are now taking advantage of this new benefit and saving money."

Let's do a little fact-checking, shall we?

At the bottom of that press release is a handy table. It shows 30 million people registered for Medicare Part D, and an additional 5.8 million people with "credible drug coverage". The 30 million people described as 'registered for Medicare Part D' are further broken down into categories:

• Stand-Alone Prescription Drug Plan (PDP): 8.1 million


• Medicare Advantage with Prescription Drugs (MA-PD):5.8 million


• Medicare- Medicaid (Automatically Enrolled): 5.8 million


• Medicare Retiree Drug Subsidy (RDS): 6.8 million


• FEHB Retiree Coverage: 1.6 million


• TRICARE Retiree Coverage: 1.9 million


• TOTAL: 30.0 million

The press release claims that the administration projected that Medicare Part D would have 28-30 million "enrollees" by May 15. But if you check the original projections, it turns out that the administration has changed its definitions. In 2005 (warning: huge html page), 'enrolling in a Medicare Part D plan' meant enrolling in a stand-alone prescription drug plan (PDP) or a Medicare Advantage plan (MA-PD). (Dual eligibles are counted in this number; they are enrolled in a PDP automatically.) This means that in 2005, the government counted only the first three groups in the table above as "registered for Medicare Part D". Now, it counts all six groups*.

Likewise, in 2005, 'people receiving creditable drug coverage' meant: these three groups plus the fourth group (those who get prescription drug coverage through an employer plan that gets a subsidy from Medicare.) Now, "credible drug coverage" includes all sorts of people who weren't included in 2005 at all.

If you hold the categories constant, here's what you find:

In 2005, the administration projected that there would be 29.3 million people in the first three groups in the table. As of now, there are 19.7 million. We're 9.6 million people short in this category.

In 2005, the administration projected that there would be 39.1 million people in the first four groups in the table. As of now, there are 26.5 million people in those four groups. We're 12.6 million people short here.

We are, in short, not meeting our projections, despite what Secretary Leavitt says.

It's actually even worse than it looks. Recall that the number of people actually enrolled in Medicare Part D, using the more honest 2005 definition, includes three groups: "dual eligibles", people in Medicare Advantage programs, and people who sign up for stand-alone Part D plans. Of these, dual eligibles were enrolled automatically at the beginning of the year. (You may remember the headaches they encountered.) There was absolutely no question about whether they would enroll: they had to, whether they wanted to or not. All of them are already enrolled.

People with Medicare Advantage plans have, in general, a much easier enrollment process than others. They already have a plan; it offers them prescription drug coverage; they take it. See how easy? As a result, a lot of them have already enrolled. In fact, according to the Kaiser Family Foundation (pdf), there were only 4.8 million people with MA plans as of Feb. 13, 2006; since the government claims that 5.8 million of them have signed up for Part D coverage, it's probably fair to say that there are not a lot of people with MA plans waiting to enroll. (No, that's not a typo: the government says that a million more people have signed up for Part D through their Medicare Advantage plans than actually had such plans a couple of months ago. It does make me wonder whether one of the figures in the government's press release is a typo, though: it seems odd that the number of people in MA-PD plans and the number of dual eligibles should be exactly the same.)

This means two things. First, we have already got all the low-hanging fruit: people who were either enrolled automatically, or for whom enrolling was easy. There are no more dual eligibles, and very few more people with MA plans. Second, this means that those missing 9.6 people will have to come overwhelmingly from those people who are not currently signed up for a Medicare Advantage program, and who therefore have to navigate the bewildering process of choosing a prescription drug plan. Only 8.6 million people have signed up this way so far; we need to more than double that figure in the next three weeks if we are to meet the government's projections.

***

At this point you might be asking yourself: so what? Sure, it's wrong for Secretary Leavitt to inflate his statistics to make it seem as though things are going better than they really are; but does it really matter whether or not the government meets its projections?

Unfortunately, the answer is: yes. And the reason why has to do with risk pools.

The insurers who participate in Medicare Part D are, naturally, hoping to make money. They will make money only if the money they take in from premiums, plus subsidies from the government, exceeds what they pay. And that will be true only if their average cost, per person, is reasonably low.

One would expect that the people most likely to sign up for a prescription drug coverage plan would be those with high prescription drug costs. And in this case, one's expectations are right, at least according to this poll (h/t Kate Steadman, again), those who do not plan to get Part D coverage "are lowest users [sic] of medical services. Those who said they had not chosen a plan and had no plans to do so take fewer prescriptions; spend less on prescriptions; go to the doctor less often; and make fewer ER, inpatient hospital and outpatient clinic / surgery center visits." If those people stay away, and only those who spend more on prescriptions sign up, it will be a lot harder for insurance companies to make money by offering Medicare prescription drug plans.

The more people sign up, the less likely it is that they will be, on average, a lot sicker and more costly to cover than your average senior citizen; and the more likely the plans are to be financially sustainable. The fewer people sign up, the more likely it is that premiums will go through the roof, and that the program will, at some point, implode.

Which is to say: this does matter, and the fact that we need, somehow, to find nearly ten million new enrollees in the next three weeks is not exactly encouraging.

***

* Wonky footnote: From the Medicare Part D Final Rule, pp. 4458-4464, here are the government's projections for 2006, as of Jan. 2005, which is when it promulgated these covering regulations (the most useful tables are on p. 4460):

Total number of people eligible for Medicare Part D: 43 million.

"Total Beneficiaries Estimated to Receive Creditable Drug Coverage, Either Through Medicare Part D Plans (PDPs or MA-PDs), or Through Employer/Union Sponsored Retiree Plans That Are Eligible For the Medicare Drug Subsidy": 39.1 million.

"Total Beneficiaries Enrolling in Medicare Part D Plans": 29.3 million.

Total number of dual-eligibles (included in the 29.3 million Part D enrollees): 6.3 million.

I have put some of the definitions in quotes, rather than summarizing them, so that you can see exactly what the administration was projecting these numbers for, rather than having to take my word for it.

Comments

[Tim]

Hilzoy, I really think that you should reduce the stress on yourself and only document the instances when the Bush admin has told the truth about something. You can then use your copious free time for house posts.

[kenB]

As far as the projected enrollments for the year go, it's really too soon to say -- facilitated enrollments have just started (for moderately low-income beneficiaries), and the late enrollment penalty doesn't kick in until the summer. Which is not to say that their press release isn't misleading, however.

BTW, it is indeed "creditable" coverage and not "credible" coverage, although the latter is a common mistake.

Re the discrepancies in the MA vs MA-PD numbers, Feb 2006 isn't a good reference point, since by then a large number of MA members had already gone to an MA-PD (in fact many were automatically rolled over into an MA-PD -- they would have had to take positive action to opt out. This is another fact which is glossed over in the press release, along with the fact that a good chunk of that population already had drug coverage through their MA plan). The real reference point you want is the MA membership in December 2005 or prior.

[hilzoy]

kenB: Yes, I saw the credible/creditable thingo. 'Credible' is from the HHS press release. In the official enrollment stats, however, it's 'creditable' (but I think you have to download the Excel file to see it.)

I think the late enrollment penalty kicks in as of May 15, which is, horrifying to say, just 3 weeks away. So I'd expect a surge of people between now and then, and a lot fewer thereafter. However, I think it's unlikely that that surge will give us nearly ten million new beneficiaries.

[kenB]

Re "creditable", yes, I wasn't correcting you, just confirming which one was correct. 80% of the people in the biz, whether at CMS or in the health industry, say "credible" instead -- apparently it's just too much work putting that fourth syllable in the middle, nevermind that "credible" doesn't really make sense in that context.

Anecdotally, I've talked to several (healthy) seniors who've been putting off their decision as long as possible, until there were nontrivial negative consequences to waiting. The penalty works out to 32 cents extra per month times the number of months the member delayed enrolling (at least for this year, probably comparable in future years), so it's possible that different people will have different thresholds for what's too much to pay in perpetuity. Whether there are 10 million people out there waiting for just the right moment or not, I have no idea. I also have no idea how many Medicare benes are even aware of the penalty.

As for when the penalty starts, I remember from last year that there was some talk about that not kicking in until August, but I haven't heard anything more about it recently. Also, I don't remember if it was just the extra charge that was going to be started in August or that they'd be pushing back the actual reference date for determining the penalty. Probably just the former -- the pushback was probably due to the schedule for the required systems changes, not to any desire to give seniors a break.

[javelina]

I'm just trying to figure out what (if anything) to suggest to my parents. So far have come up very blank indeed.

[John Lopresti]

There is a political party's website that covers the same topic on which Hilzoy has done so much work; that other site is here.
We remember what the administration did during three years preceding the attempted rollout of "private social security accounts", suppressing Richard Foster's actuarial work which revealed the weaknesses of the bookkeeping involved in that proposed law; here is an article written about that debacle two years ago, mentioning his ultimately getting fired.
Ostensibly, if the accounting is the difficulty here, as Hilzoy has highlighted, all that might be needed, once again, would be to locate, suppress, and then fire the actuarial people who examine the bookkeeping; that might be do-able in the next three weeks, to meet the target date.
Credible is the byword of this administration; as are elided syllables. Yet, to be fair, I work with many people with advanced degrees, some of them with a similar kind of neo-vernacularization of words, as contributors here discussed. It might be appropriate to say this administration has a creditability problem in healthcare matters, though, usually the criticism would be of its credibility.