John Graves of Vanderbilt’s health policy group raises an interesting point that I want to expand on:
A uniquely American dynamic to keep tabs on is that as the end of insurance plan years approaches, its well known that people schedule ($$$) elective medical procedures before their deductible resets on Jan 1.
— John Graves (@johngraves9) October 2, 2020
We know that massive amounts of deferrable care was pushed off in March, April and May. Utilization is still down for the year as people aren’t going to the hospital as readily or as quickly even as there are beds open and hospitals are performing a near normal surgery and procedure schedule.
We also know that the 4th Quarter of the year tends to be fairly heavy on elective procedures. The financial incentives are really clear here.
Some people will know that they will hit their deductible and out of pocket spending maximum every year. They know that they are going to have to come up with the cash again in January-February-March of next year. Maxing out their deductibles and out of pocket limits does not change their expected future spending and scheduling a procedure in November instead of March might be a cash flow decision, it might be a decision that the doc that they like is available, it might be a decision that the chunk of cartilage floating in their knee hurts like a son of a bitch or it might be a decision that they have a vacation scheduled in March and want to be up and about for all of it. There are good reasons for people who know that they are always going to max out their spending limits to schedule in one year or another, but the timing won’t change the two year total spending.
However, most people don’t consistently max out their out of pocket spending. Most years, most people barely interact with the medical system. However in some years, some people will max out their cost-sharing. They could have had a baby. They could have broken a leg and needed surgery. They could have had a set of symptoms that required expensive imaging to rule out really bad outcomes and have the symptoms resolve on their own. Who knows what the story would be. But the important part is that these individuals would have a reasonable expectation that the next year won’t be a max-out year.
In that case, timing matters. If there are issues that have been deferred and could be deferred for a while long, timing really matters. A sleep study performed in November and a CPAP machine arriving in December could cost the patient thousands of dollars less than the same exact service and delivery in January. A minor surgery to clean up bone fragments from an untreated ankle injury incurred while refereeing a 2013 soccer game could cost nothing if done in October and $3,000 if done in February.
People who maxed out their deductibles and have a reason to believe that they won’t have large medical expenses in the next contract period will shift services from the future where cost-sharing applies, to the present where cost-sharing is not relevant for a little while.
We know this happens. Actuaries expect it to happen. And actuaries can incorporate past behavior into future projections that account for this behavior. However, it is extremely likely that in 2020 the number of people who are going into the 4th Quarter of the year fully maxed out is less then normal. It is also quite plausible that of the people who are maxed out, the probability of scheduling another procedure is lower than it otherwise would be.
If we assume those two things are true, why does this matter?
In the short run, insurers will make more money as claims will be lower than premiums by a bigger margin than expected. Some of the procedures that we would have expected to happen in November 2020 will occur at some point in 2021 but not all. More importantly, it makes the ability to project the future fuzzier. Actuaries count on the near future to look a lot like the near past. The shock to utilization patterns due to COVID is massive and makes the ability to project 2022 or 2023 from 2020 data very questionable. Projections will either rely on older data, fuzzier data or shorter data spans. If that is the case, good faith professional estimates are likely to have wider error bands for several years going forward until whatever the new normal is stabilized.