Two weeks before Independence Day, the United States had a legitimate reason to celebrate. After more than a year of worry and restrictions that followed the emergence of the coronavirus, the number of new cases each day had fallen to a little over 11,000, the lowest level since the virus first emerged in March 2020. The rapid vaccination of millions of Americans had cut off potential hosts. Combined with warmer weather that pushed people outside, the virus was far less able to gain ground.

Over the past two weeks, though, the picture has been different. Thanks in part to the spread of the delta variant, the number of cases in the United States has shifted upward. The average number of new cases at this point is 76 percent higher than it was two weeks ago, the biggest increase since last July. Following the pattern we’ve seen over the past year, hospitalizations have also increased.

The urgent question now becomes: Will the number of deaths increase, too?

There are two immediate ways in which one might expect that question to be answered. The first is that, yes, an increase in cases necessarily means an increase in deaths, because any infection bears at least some risk of that outcome. But the second answer is that the number of deaths should be lower than in months past, since the vaccines have proved effective at muting the worst effects of covid-19, the disease caused by the virus. But there is reason to worry that the latter answer includes some dangerous loopholes.

There are four pertinent metrics here: the number of new cases (expressed as a rolling average over the prior seven days), the number of people currently hospitalized, the number of new deaths and the number of vaccinations. Since the beginning of the year, here’s how each metric has shifted.

Each of those charts includes an important detail. The uptick on cases. The smaller uptick on hospitalizations. The relatively flat number of deaths. And, critically, the surge and fade of new vaccinations.

What’s important in answering our question is the level of change for each metric. If we look at how each has shifted relative to two weeks prior, you see that cases are increasing quickly, with hospitalizations starting to catch up. The number of deaths is still down from two weeks ago, as, unfortunately, is the number of vaccinations.

The question, really, is where the deaths data will go next. Over the course of the past 10 months, there’s been a fairly consistent pattern: New cases lead to new hospitalizations a week later and new deaths two weeks after that. If we look at the change in each metric with hospitalizations and deaths offset from the new case totals, you can see the wavelike pattern that the three numbers have followed.

That black line, the number of deaths, remains flat — but is also at the point where history would suggest it’s going to start heading upward.

Theoretically, the fact that so many older Americans are vaccinated should prevent a big spike from occurring. After all, 89 percent of Americans over the age of 65 have received a dose of a vaccine, with 79 percent being fully vaccinated. And the vast, vast majority of those who have died of covid-19 are the elderly.

But that hospitalizations are increasing about in line with what we’d expect is worrisome. Most of those getting sick are unvaccinated, as are nearly all of those — 99.2 percent — who are dying. But because the density of vaccinations in many parts of the country is not robust enough to shut down the spread of the virus, there remain a lot of people at risk of contracting the disease.

In some cases, that rejection of vaccines is willful. Polling from The Washington Post and our partners at ABC News last month found that more than 1 in 10 people over age 65 say they won’t get vaccinated, creating a pool of about 7 million Americans in the highest-risk group who refuse to protect themselves.

It’s also important to point out that even in May, by which point vaccines had been widely available to the oldest Americans, most of those dying from the virus were 65 and older.

Data from the Centers for Disease Control and Prevention come at a significant delay, so it’s hard to say with authority what more recent national figures would look like. But this doesn’t suggest a significant break from the pre-vaccine pattern on deaths.

It may be the case that the relationship between cases and deaths will shift as a result of the protection offered by vaccines. From September to December, there were about 1.75 deaths for every case four weeks prior. So far this year, the ratio has averaged about 1.3, a significant improvement. There’s been a similar drop in the ratio of hospitalizations to cases. Both of those ratios are currently increasing, as a function of the low number of new cases a month ago.

Will those lower ratios hold? Will they drop further? Will we see far fewer deaths for every new case in six months’ time? Or are we going to repeat the same pattern, with nearly as many deaths for every new case even as those who are vaccinated are overwhelmingly protected? Will the delta variant itself skew that calculus?

The unfortunate answer to those questions is: We will apparently soon find out.