David Anderson

Overly-optimistic actuaries are a good way for an ACA individual market insurer to lose money.

Actuaries and benefit managers are trying to price out their ACA plan offerings for 2021 right now.  CMS requires initial rate filings to be made by the start of June with final rate submissions at the end of July.  Some states have earlier deadlines.  Pennsylvania wants initial filings in mid-May. After insurers file an initial rate, they engage in a back and forth with their state regulators, whose primary concern is solvency, to get to a final rate. The final rate will often incorporate new information and claims experience that has emerged in the several months between initial filing and final approval.

Actuaries have a process to generate a rate filing.  They use the most recent year of complete data, in this case 2019, and then apply a model that accounts for changes in the population, changes in network and technology, changes in payment levels and changes in underlying health status to come up with a rate that is not inherently banana-pants.  This process works reasonably well when the recent past can reasonably be assumed to be a decent approximation of the near future.

The ACA had an optimistic actuary problem in 2014.

No actuary in 2013 really had a good idea what the 2014 guaranteed issued, community rated, subsidized individual health insurance market would look like.  The price linked subsidy system of the ACA where the benchmark premium is locally determined by the premium of the second cheapest silver plan means that every insurer was competing to be the cheapest silver plan as the insurer with the cheapest silver plan determines what the second cheapest plan is.

Actuaries were pricing off of consultant reports and guesses at the underlying health status and premium sensitivities of the covered population.  Some insurers thought that the market would look like COBRA.  Others thought it would look like Medicaid, and more thought it would look like large group, commercially insured.  Since the actuaries had no clue what the 2014 population would look like, they took educated guesses. And if we assume that most insurers have fundamentally similar enough cost structures with the exclusion of the Medicaid Managed Care insurers like Molina and Centene/Ambetter, the insurer with the most optimistic guess as to the relative health of the population would price lowest and “win” the price war.  That “win” led to a low price level in 2014 and 2015 that was significantly below claims costs.  It was a failure of a first price auction where the most optimistic predictors won marketshare by pricing too low to cover claims costs.

The ACA may have an optimistic actuary problem in 2021.

Right now, there is massive uncertainty as to what utilization and health profiles will look like in 2021.  We can assume that individuals who have been on ventilators in 2020 will need continued, elevated levels of care in 2021.  We can assume that significant number of procedures that either have already been cancelled or will be cancelled as hospitals make space for COVID-19 treatment will eventually be performed in 2021 or 2022.  We can assume that some of the regulatory steps that are greatly expanding the licensed clinical workforce such as full scope of practice laws for master-level clinicians and far more telemedicine/distance medicine, will both continue.

How many people will have COVID-19 and be ventilated in 2020 and need extended follow-up care in 2021?

How many people will have COVID-19 in 2021 and be hospitalized?

What does a provider network look like in 2021 after all of the clinicians have been infected with COVID-19 in 2020?

How many procedures that were deferred in 2020 will be people try to schedule for 2021?

How does telemedicine and master level clinicians bill in 2021?

All of those are among the dozens of known uncertainties.  Experts can reasonably choose different points on all of these questions as to what the plausible future of 2021 looks like.  Some experts will have an almost random distribution of what they think 2021 will look like on these questions.  Some experts will be uniformly more pessimistic or optimistic than average on each question.

The optimistic actuary who thinks that there won’t be a ton of COVID-19 2020 cases that need follow-on care in 2021 and that the provider networks will be as broad or broader with more low cost master level clinicians who can perform most services on an outpatient, non-facility fee basis while most deferrable procedures won’t be rescheduled will produce a very different premium than an actuary who thinks that there are a ton of people leaving the ICU who will have long lasting and costly health shadows of the future while also expecting hot spots to break out in 2021 even as most of the deferred care comes rushing back to mostly inpatient settings in 2021.

Right now as actuaries are preparing their initial rate filings, good faith error bands are going to be huge.  I expect rate filings in May to have huge variation in initial rate increases between insurers with similar 2019 and 2020 rate structures just because Company A will be more optimistic than Company B on some measures as we won’t have any clue what is happening with COVID19.  However, I also expect rate requests to converge significantly on final approval as the extra few months of information about the peak and scale of the COVID19 experience will dramatically reduce the variance in optimistic and pessimistic actuaries.

So as rates are filed in the next six to eight weeks, expect to see some eye-popping requests and some numbers that just seem a bit low.  Those will change as we know more.

Rate setting, optimistic actuaries and pandemicsPost + Comments (2)

The New York Times has some comparative good COVID-19 news from Washington State:

While each infected person was spreading the virus to an average of 2.7 other people earlier in March, that number appears to have dropped, with one projection suggesting that it was now down to 1.4…

Washington State’s coronavirus figures have continued to grow steadily, but not as fast as other states’. The death toll has been doubling about every eight days in Washington, compared with every two or three days in New York, Michigan, New Jersey and Louisiana.

Washington State shut down hard and shut down fast compared to the rest of the country. This is what success looks like right now. The death rate is still increasing, the infection rate is still increasing, but the rates of increases are slower than the rest of the country. This is success on the second derivative.
As I mentioned in my first post this morning, most of the people who are are hospitalized this morning were infected at least two weeks ago. Washington State has been able to shrink the number of people who were exposed and infected two weeks ago with aggressive and early public health measures that were several doubling cycles ahead of many regions shutting down.

Washington State is what success could look like for the next month or two, and it is still going to be rough.

Success is the 2nd derivativePost + Comments (58)

I’ve been at Duke and in academic research for a little more than three years. There are two important categories of helpful questions that I always need to think about:

• “What exactly are you counting and are you counting what you think you are counting?”
• “What, exactly, is your denominator?

We need to ask these questions when we think about COVID-19 right now as we know two things about it. First, we are still in an exponential growth phase of infection spread where success at this point is merely slowly the exponent of the growth function in the United States at this time.

22 days from 100 cases to 100,000, because that’s how exponential growth works when left unchecked.https://t.co/Lytl2cz0Ph pic.twitter.com/RdpaRRKjqD

— Justin Wolfers (@JustinWolfers) March 27, 2020

Right now, it seems like the US is seeing a doubling of cases every two to three days. Some of this is a function of testing starting to come online picking up lower acuity cases, but a lot is probably new cases.

The other thing that we know is about COVID 19 is that there is a significant lag between infection and symptoms and then symptoms to hospitalization for the people who are hit the hardest by the disease.

Annals of Internal Medicine## estimated the time from infection to detectable symptoms was a median of 5 days.

After symptoms were detected in China, hospital admissions were usually happening a median of seven days for all hospitalized patients and eight days for patients who went to an ICU.

If we can safely add up median time from infection to symptoms and then symptoms to hospitalization, that sums to a back of the envelope span of 12 to 13 days.  In the United States, that translates to between 4 and 6 doubling cycles.  Most of the people who are presenting ICUs were infected by the time significant social distancing started in the United States.  The ACC basketball tournament was still being played in front of 8,000 fans on March 12.  New York City public schools were still open on March 15.  Spring Break and Mardi Gras were in full swing.

From a March 12, 2020 Lancet article**, a very good point is made about death rates as a function of time infected:

However, these mortality rate estimates are based on the number of deaths relative to the number of confirmed cases of infection, which is not representative of the actual death rate; patients who die on any given day were infected much earlier, and thus the denominator of the mortality rate should be the total number of patients infected at the same time as those who died…

All of this is a long way to say that I am extremely perplexed by this following tweet:

FWIW: New York now has a pretty large sample size of coronavirus cases. And, so far, the hospitalization rate is between 13 and 15 percent. The ventilator rate is under 4 percent. These are, so far!!, better than worst fears.

— Dan Goldberg (@DanCGoldberg) March 28, 2020

If this is Total Hospitalizations/Total Cases or Total Ventilation/Total Cases, then we have a huge timing problem as most cases (50%) are cases that happened in the past doubling  and 87.5% of the cases are from individuals who were infected in the past three doublings.  People who have been infected in the past two or three doubling cycles are (at this time) extremely unlikely to be already dead, already ventilated or already hospitalized.

The CDC has a good summary of the complete case outcome from Diamond Princess where there was a effectively a single infection event and significant follow-up:

Among 3,711 Diamond Princess passengers and crew, 712 (19.2%) had positive test results for SARS-CoV-2 (Figure 1). Of these, 331 (46.5%) were asymptomatic at the time of testing. Among 381 symptomatic patients, 37 (9.7%) required intensive care, and nine (1.3%) died (8). Infections also occurred among three Japanese responders, including one nurse, one quarantine officer, and one administrative officer (9). As of March 13, among 428 U.S. passengers and crew, 107 (25.0%) had positive test results for COVID-19; 11 U.S. passengers remain hospitalized in Japan (median age = 75 years), including seven in serious condition (median age = 76 years).

In that case the denominator does not have time variation and the Case Fatality Rate (CFR) is not complete as there are still people in the ICU who could quite plausibly die, but there has been enough time to establish a somewhat reliable number.

When we see reporting that is a function of division, we always need to ask what are we counting, and are we counting what we think we are counting and finally are we using the right denominator?

## DOI: 10.7326/M20-0504

** Lancet Infect Dis 2020 Published Online
March 12, 2020 https://doi.org/10.1016/S1473-3099(20)30195-X

Exponential growth and death date delaysPost + Comments (20)

In a normal year, I, and several other ACA nerds would be getting ready to start tracking ACA premium rate hikes. Initial filings are due in April, and the actuaries and plan designers have been looking at their 2019 experience with whatever information that 2020 could hint at to predict what costs and utilization would look like in 2021. If asked in the first week of January, I would expect, on average, for rates to have gone up by a fairly normal amount of inflation plus a point or two as there would have been no major policy shocks and the overpricing that happened in 2018 and 2019 is likely to have mostly worked its way out of the system with attendant large MLR rebates to be issued this fall.

HOWEVER, we’re not in normal times. COVID-19 will significantly mess up normal actuarial models.

Rates set for a future year are supposed to be sufficient to cover expected claims, administration, reserve accumulation as needed and produce a regular profit. They are not supposed to be used to make up for prior year losses. Prior year experience can be used however to inform revisions of estimates of utilization, costs and population profiles.

One of the big questions for all actuaries who are trying to figure out what type of rates to set for 2021 for any product line (Medicare Advantage, Medicaid Managed Care, Exchange, Employer Sponsored etc) is how elastic are services?

In 2020, utilization patterns were normal in January with a bit more flu than typical but it was well within normal variance, and lots of people getting their knees replaced, or shoulders repaired and the other 10,001 things that happen in hospitals. In February, normal utilization patterns were still happening. By the end of the month, some hospitals were starting to make initial plans to prepare for a COVID-19 surge. Some procedures were being diverted or cancelled. By early March, many hospitals had started to clear the decks in preperation for a crisis. My mom, the medical zebra, had four appointments scheduled for the 1st week of March. Two were cancelled outright, one became a virtual visit and the last one was an annual MRI that is needed to monitor the effects of previous surgery. This was not uncommon but not widespread. By March 20th, the hospitals that had committed to pre-planning were as empty as they were going to be as they sent everyone home who could go home. And now, the COVID-19 surges are coming. And the surges are likely to come for weeks or months.

People who need critical care will still have a chance of getting critical care. People who need care that can be provided at a distance may still get that care. People who need pharmaceutical interventions may still get that care, but a lot of people who have legitimate care needs that is not life nor death important won’t get that care right now. The hospitals are at full tilt with COVID-19 and will be at full tilt for weeks or months.

The actuaries need to figure out, guestimate, project, open up and appropriately read goat entrails, how much utilization that normally would have happened this year and did not will happen next year.

This calculation will vary. Acute care for sprained ankles that is now being treated by a phone call and a recommendation to rest, ice, compress and elevate in May 2020 instead of an urgent care visit won’t show up in a January 2021 urgent care claim. Someone who would have had their knee replaced in July in a normal year is likely (assuming they are alive and healthy) have it replaced sometime in 2021 instead. Actuaries have to figure out how elastic subsets of claims are and project that into next year’s premiums. This is challenging as there is some literature on this but always on a far smaller scale. In the ACA context, the most optimistic actuary is likely to price too low even as the optimistic actuary is what will drive and shape the market.

Rate setting for 2021 and claims elasticityPost + Comments (2)

Early this week, I thought that Medicaid and the ACA Exchanges will both be doing an incredible amount of heavy lifting on maintaining and expanding health insurance and health finance coverage during the pandemic. Unemployment is going to be massive and COBRA will be too expensive for many people.

Unemployment Insurance Weekly Claims

Initial claims were 3,283,000 for the week ending 3/21 (+3,001,000).

Insured unemployment was 1,803,000 for the week ending 3/14 (+101,000).https://t.co/ys7Eg5LKAW

— US Labor Department (@USDOL) March 26, 2020

These are high deductible plans. The cheapest plan in most counties will usually have a deductible over $7,000. That is a lot. However, the choice is a zero premium, capped deductible plan with some pre-deductible services including COVID19 testing in a pandemic or a zero premium, infinite deductible plan.

If you are currently uninsured and are in one the SBM states, look for a zero premium plan at the very least during this special enrollment period.

Special Enrollment Periods in State Based MarketplacesPost + Comments (6)

The Senate passed the House Covid19 paid sick leave and Medicaid payment bump bill 90-8 yesterday.

BREAKING: Senate passes House coronavirus aid package 90-8.

Senate GOP inches closer to unveiling their proposal for a third, larger stimulus package to address the epidemic, and Treasury sends topline figures to the Hill for individual payments & loans.https://t.co/P3QzlXZVCU

— Andrew Desiderio (@AndrewDesiderio) March 18, 2020

This bill, once it is signed into law, will have two major effects. First, it will give states a significant cash infusion routed through the Medicaid program. State sales and income tax revenue is crashing this week and will be at minimal levels going forward. 49 out of 50 states have balanced budget constraints. So a big infusion of federal funds to cover more of the Medicaid expense will mean states won’t be cutting back as deeply as they otherwise would have been.

Secondly, the paid leave portion of the bill is the first downpayment that we, as a society, are making to pay for externalities. The public health measure of physical distancing requires people to act, at high cost to themselves, for the good of everyone else. This ask is an ask for impoverishment for the vast majority of the working population. People can’t afford to not go to work as they have rent or a mortgage to pay, food to buy and lights to keep on. With paid leave for some that will (eventually) be fully paid for by the federal funds, we’re making it easier for some more people to stay home and break infection chains.

The 3rd round stimulus package has a possibility of being designed as a massive purchase of more externalities. There is also a set of designs where it is just a redistribution of public wealth to well connected private interests. But the current second package that is now heading for a signature is a large, first attempt at buying some needed externalities.

We’re paying for externalities nowPost + Comments (116)