How realistic are my final week projections?
2018 MIDTERM ELECTION
Time: D H M S
So, I was dead-on target for HC.gov enrollments as of February 6th. What does that mean for the final 9 days (which we're halfway through already)?
Here's my final (national) projections again:
- Mon. 2/09: 10.57 million (90K)
- Tue. 2/10: 10.68 million (110K)
- Wed. 2/11: 10.81 million (130K)
- Thu. 2/12: 11.0 million (190K)
- Fri. 2/13: 11.3 million (300K)
- Sat. 2/14: 11.8 million (500K)
- Sun. 2/15: 12.5 million (700K)
There's two issues here: First, is the demand going to be there? Second, can the system handle the strain if it's there?
According to Andy Slavitt of CMS:
Right now, we're focused on the next few days on being there for our consumers and to help them select the best plans for their families and the final days of open enrollment. As expected over the last couple of days, interest in signing up for coverage in the final week of open enrollment is beginning to increase and we're seeing a really nice ramp.
For example, traffic on HealthCare.gov has been increasing over the course of the week. As of 10 am, there was a 58 percent increase in visitors to the marketplace compared to 1 week ago last Wednesday.
The volume of calls coming into the marketplace call center is also showing a marked increase from this time last week with about 37 percent more calls week-over-week. Despite the increase in volume, both the Web site and the call center are operating well as expected with minimal wait times.
OK, obviously increased visitors and calls doesn't necessarily correspond 1:1 to increased enrollments, but let's suppose that it does. A flat 60% increase over last week would mean 440K QHPs via HealthCare.Gov itself, and likely about 570K nationally, as of Friday the 13th. That would bring the grand total up to only around 10.7 million, plus whatever comes in over the final 2 days.
I, however, think that we've already hit 10.8 million today, and will reach 11.3 million by Friday. That would require about 1 million even nationally from 2/07 - 2/13, or 143K/day nationally. On HC.gov, I'm projecting a cool 750K from 2/07 - 2/13, or 107K/day...a 2.7x increase over last week.
That's not a problem system-wise, assuming the demand is there.
The real question is Saturday and Sunday (not just a weekend, but Valentine's Day weekend, no less). Will things look like they did last spring? Will it be a bit overhyped, with a nice-but-not-amazing surge?
Assuming I'm correct, we should be looking at roughly 1.2 million people nationally in just 2 days, with around 900K of them via HC.gov. HHS has stated that the site is supposed to be able to handle up to 250K concurrent users; assuming 1 hour per session, that's theoretically up to 6 million people in a single 24-hour period, so 450K/day shouldn't be a problem. Of course, that assumes an even spread including the middle of the night, which isn't likely, and some of the state-based exchanges which have still been having technical issues (I'm especially concerned about Washington State) may not be able to handle it.
Of course, if there are serious technical issues at HC.gov or other exchanges, one or more of them may indeed go ahead and announce an extension period, either "full" or of the "as long as you started the process by midnight..." variety.
In short, I'm pretty confident that the system can handle a massive final surge...the only question now is whether the demand will be there for it.
Here's another way of looking at it: What enrollment data is missing as of today? Well, let's assume that the missing days were to only equal the daily average so far. That would mean:
- 37 HC.gov States: 7,749,375 / 84 days = 92,254/day x 9 days = 830.3K
- California: 1,232,568 / 80 days = 15,407 x 13 days = 200.3K
- Colorado: 125,378 / 78 days = 1,607 x 15 days = 21.1K
- Connecticut: 99,000 / 86 days = 1,151 x 7 days = 8,100
- DC: 19,837 / 66 days = 300 x 27 days = 8,100
- Hawaii: 16,368 / 63 days = 260 x 30 days = 7,800
- Idaho: 90,567 / 64 days = 1,415 x 29 days = 41K
- Kentucky: 95,927 / 83 days = 1,156 x 10 days = 11,500
- Maryland: 100,547 / 82 days = 1,226 x 11 days = 13.5K
- Massachusetts: 122,600 / 88 days = 1,393 x 5 days = 7,000
- Minnesota: 45,973 / 83 days = 554 x 10 days = 5,500
- New York: 453,968 / 81 days = 5,604 x 12 days = 67.3K
- Rhode Island: 30,669 / 85 days = 361 x 8 days = 2,900
- Vermont: 29,055 / 87 days = 334 x 6 days = 2,000
- Washington:158,000 / 88 days = 1,795 x 5 days = 9,000
Add all of those up and you get 1,235,400 added to the 10,359,832 confirmed QHP selections, for a total of 11,595,232 (call it 11.6 million).
I'd call 11.6 million the baseline at this point; that is, I'd be very surprised if the total comes in below that. Multiply each of the above by about 1.7x and you're at 12.5 million.
UPDATE: Here's a third way of looking at it. I went back to my 2014 spreadsheet, and as far as I can tell, it looks like total QHPs crossed the 5.4 million mark at midnight on March 24th. They hit 7.1 million as of midnight on March 31st. That means they managed to enroll roughly 1.7 million in the final 7 days of the official enrollment period. My projection assumes around 2 million in the final week of 2015's open enrollment period, which isn't all that much higher (I was actually surprised by this; I hadn't checked those particular numbers when I came up with my 2015 "curve")