What does it look like when Obamacare explodes? This interactive graphic explains.

Obamacare’s insurance marketplaces are in a precarious position.

Aetna quit the Obamacare marketplaces last week, leaving Nebraska and Delaware.

In Iowa, a smaller plan called Medica has threatened to follow suit. If it quits, 94 of Iowa’s 99 counties would have no Obamacare plan at all.

Insurers leaving the marketplace have cited growing instability and uncertainty over the health care law’s future. Some previously lost millions in the marketplaces’ early years, when they had more sick people enroll than expected. Health plans worry that Trump administration actions could exacerbate those problems.

The biggest risk when health insurance plans quit Obamacare is that some areas could end up with no plans at all. This would mean that, while the law was still technically standing, people there would not have access to the program. They’d have no place to use the financial help the government provides to buy coverage on Healthcare.gov.

To see what areas of the country are most at risk for this type of problem, we worked with Kathy Hempstead at the Robert Wood Johnson Foundation to visualize the effect of four major Obamacare plans leaving the network. We chose Anthem, Molina, Centene, and Cigna because they all have a big footprint in the individual market.

If all four of those insurance plans left, there would be 241 counties in the United States with zero plans selling individual market coverage.

The areas most at risk for this type of collapse are largely also places that voted for Trump: places that are lower-income and rural, which aren’t attractive markets to health insurance companies. These are places that have struggled to attract robust insurance competition both before and after the health care law’s implementation.

The Obama administration worked hard to recruit health insurers to sell to those areas. The Trump administration, however, seems to want to stand aside and let Obamacare run on autopilot, so it can explode or survive on its own. We can now see where the potential for “explosion” is the highest, by looking at the footprint of each health plan — and what their exit from the marketplace would mean.

Kentucky has had one of the country’s most successful Obamacare coverage expansions. Its uninsured rate has fallen from 20.4 percent in 2013 to 7.8 percent in 2016. An Anthem exit could threaten all that, leaving 59 counties in the state without an available health plan.

Missouri would be another place that would struggle in the wake of an Anthem exit, with 85 counties in the central and southern areas of the state left without any individual market plans.

Anthem has so far been cagey about its plans for 2018 and whether it will continue participating in the marketplaces. Reports earlier this spring from analysts who follow the company had said it was “leaning towards exiting.” But Anthem has decided to stick with a few marketplaces. We know that the insurer currently plans to participate in the Virginia and Connecticut marketplaces, which have especially early rate filing deadlines. Kentucky does not require insurers to submit their rates until June 7, so we are still waiting to see what decision Anthem will make in that state.

Health insurers that have experience selling Medicaid have generally done better in the Obamacare marketplaces. They tend to have experience designing plans for the lower-income population buying coverage.

This has certainly been the story with Centene, a St. Louis-based Medicaid managed care company. It is not one of the better-known health insurance plans, but it has quickly become a dominant plan on the Obamacare marketplaces.

Last year, it had over 1 million Obamacare enrollees. It turned a profit on those enrollees, too.

Right now, Centene says they’re committed to sticking with the health law’s marketplaces in 2018 — and that is especially good news for Mississippi. Centene is the only health plan that serves the state, which currently has 88,000 marketplaces enrollees.

Centene also sells coverage in eight other states, although it is not the only health plan in any of those areas.

Tennessee has already had a tumultuous time with its Obamacare marketplace. Humana announced in early 2017 it would quit the state’s marketplace, leaving 16 counties in the Knoxville area with zero plans next year. I took a trip to that area of the state in April, and you can read more about that here.

But just last week, Tennessee got some good news: the state’s Blue Cross Blue Shield plan would offer coverage in the Knoxville area, swooping in to save the day.

Still, this doesn’t mean that Tennessee is in the clear. If Cigna were to leave the Obamacare marketplaces, that would leave 14 Tennessee counties with zero options.

Cigna was, up until last year, quite bullish on the Obamacare marketplaces. Last summer, the health plan doubled the number of states where it sells Obamacare coverage.

But this year, things are markedly different and, as of February, Cigna was still undecided on whether it would sell Obamacare coverage in 2018. Chief executive David Cordani described the marketplaces as “fragile at best” on a call with investors.

Molina is another health insurer that is not a household name; before Obamacare, most of its business was in helping states run Medicaid programs. Like Centene, it has also found a strong foothold in the health law’s new marketplaces. Molina reported covering 568,000 Obamacare enrollees by late 2016.

Molina is a big Obamacare health plan — and arguably the most vocal about the possibility of quitting. Former chief executive Mario Molina sent Congress a letter in late April saying that the company would “not participate in the 2018 marketplace” if the government does not fund the law’s cost-sharing reduction subsidies.

Those subsidies did not get appropriated in this year’s budget bill, and Trump continues to waffle on whether he will make those payments, putting Molina’s participation at risk.

Molina currently sells coverage in 139 counties across the country. Most of those places have other insurance plans, too. There is actually just one county in the country that would have zero options if Molina quit: Menominee County, Wisconsin.

Molina is at the moment a bit a a wild card. It did send that strongly-worded letter to Congress — but then, a few weeks later, its chief executive Mario Molina was abruptly fired from his job leading the company, which had been founded by his father. So its hard to know what will happen with Molina’s future, in the marketplaces or otherwise.

The Affordable Care Act depends on insurance plans volunteering to sell coverage on marketplaces. It doesn’t have a strong back-up plan if there are certain areas where no health insurer comes forward.

“If there were any [backup plans], nobody told us about it,” says Kevin Counihan, who served as chief executive of Healthcare.gov under Obama. “I sure wish there were some.”

Counihan and his team ended up spending two months last summer crisscrossing the country, checking in with health insurers and regulators to ensure all 3,007 counties would have at least one option. And this wasn’t easy; insurers did not warm to the marketplaces as the law’s drafters and boosters had expected.

The administration was able to ensure, during the four years of Obamacare’s coverage expansion, that no county’s marketplace ended up with bare shelves.

President Trump has given little reassurance to health plans that might be on the fence about entering Tennessee or other markets. His administration has waffled on whether it will continue funding key Obamacare payments. He has said he expects the marketplaces to “explode” on their own.

“Insurers are fleeing the marketplace because Obamacare is fundamentally flawed,” Health and Human Services spokesperson Alleigh Marré said in a statement. “As more Americans are presented with higher healthcare costs and fewer options for coverage, repealing and replacing the law remains the best option. Administrative action alone cannot fix the problems Obamacare created.”

Correction: An early version of this article did not include the plans that Anthem sells in Missouri. This has now been fixed.