AstraZeneca insists that its COVID-19 vaccine is strongly effective even after counting additional illnesses in its disputed U.S. study, the latest in an extraordinary public rift with American officials.

In a late-night news release Wednesday, the drugmaker said it had recalculated data from that study and concluded the vaccine is 76% effective in preventing symptomatic COVID-19, instead of the 79% it had reported earlier in the week.

Just a day earlier, an independent panel that oversees the study had accused AstraZeneca of cherry-picking data to tout the protection offered by its vaccine. The panel, in a harsh letter to the company and to U.S. health leaders, said the company had left out some COVID-19 cases that occurred in the study, a move that could erode trust in the science.

Some experts said the new data provided by AstraZeneca was “reassuring” and that the information was likely solid enough for U.S. regulators to authorize the vaccine.

“AstraZeneca may have just been too hasty in submitting the earlier, incomplete interim analysis rather than waiting to analyze and submit the full dataset,” said Julian Tang, a virologist at the University of Leicester who was not connected to the research. He said the updated details didn’t look substantially different from what was published earlier this week.

Data disputes during ongoing studies typically remain confidential but in an unusual step, the National Institutes of Health publicly called on AstraZeneca to fix the discrepancy.

AstraZeneca had been counting on findings from a predominantly U.S. study of 32,000 people to help rebuild confidence in a vaccine that, despite being widely used in Britain, Europe and other countries, has had a troubled rollout. Previous studies have turned up inconsistent data about its effectiveness, and then last week a scare about blood clots had some countries temporarily pausing inoculations. Most have since restarted after the European Medicines Agency said the vaccine doesn’t increase the overall incidence of blood clots. But it could not rule out a connection to some rare blood clots and issued new advice for monitoring those who get vaccinated.

Now the question is whether the company’s newest calculations can end the tension. Even before the latest spat, experts had expressed concern that missteps in the vaccine’s rollout could undermine confidence in the slot, which is crucial to global efforts to end the coronavirus pandemic since it is cheap, easy to store and a pillar of the COVAX initiative aimed at bringing vaccines to low- and middle-income countries.

Earlier Wednesday, Dr. Anthony Fauci, the top U.S. infectious disease expert, told reporters he hoped that when all the data was publicly vetted by federal regulators, it would dispel any hesitancy caused by the spat. He predicted it would “turn out to be a good vaccine.”

AstraZeneca’s newest calculations were based on 190 COVID-19 cases that occurred during the study, 49 more than it had included earlier in the week. The vaccine appears especially protective against the worst outcomes, with no severe illnesses or hospitalizations among vaccinated study volunteers compared to eight severe cases among those given dummy shots, the company said. It didn’t provide a breakdown of the rest of the cases.