If there were a crisis line for corporate self-harm, AstraZeneca ought to be reported: the company can't seem to stop hacking away at its own credibility.

Every time the Anglo-Swedish pharma giant has seemed on the verge of a major triumph in its COVID-19 vaccine drive, it has managed to turn glory into ignominy, committing a series of largely self-inflicted blunders, or worse.

The latest example occurred on March 22 when AstraZeneca announced the long-awaited results of its large clinical trial conducted in the U.S. as well as Peru and Chile. The company said that preliminary analysis of its results from the 32,500 person-study showed its vaccine was 79% effective in preventing coronavirus infections.

But in the early morning hours of March 23, the U.S. National Institute of Allergy and Infectious Diseases (NIAID) issued a highly-unusual statement saying that AstraZeneca may have used “outdated information” that “provided an incomplete view of the efficacy data.” It said it took this action after an independent panel of medical experts that works under the National Institutes of Health to help monitor the AstraZeneca trial informed the agency that it had problems with the numbers the company had used in its press release on March 22.

The language used by the independent panel, known as the Data and Safety Monitoring Board, was stark, according to a copy of the letter the board sent the company which The Washington Post obtained. “The point that is clear to the board is that the [vaccine efficacy number] . . . they chose to release was the most favorable for the study as opposed to the most recent and most complete. Decisions like this are what erode public trust in the scientific process,” the DSMB said according to the newspaper.

The letter indicated that the more up-to-date data showed an efficacy of between 69% and 74%, and that AstraZeneca knew this and yet chose to go with the older 79% figure.

The company says that the numbers used in its announcement were based on a “pre-specified” cut-off date of February 17. It also said that more up to date results were “consistent” with what they announced Monday. It said it was “now completing the validation of the statistical analysis” and was talking to the DSMB to share its analysis of this more up-to-date data with them, with the results to be published within 48 hours.

Some commentators have been quick to condemn the company, saying the latest snafu is inexcusable. Eric Topol, a cardiologist who is a prominent commentator on medical affairs, took to Twitter to call AstraZeneca's response to NIAID's statement "unacceptable," saying they clearly knew what the more current data indicated and that "it should not take 48 hours to sort out." Anthony Fauci, the head of NIAID, told The Washington Post that he was "shocked" by the DSMB's concerns. “The irony of this is that it’s very likely a very good vaccine, and this sort of thing does nothing but cloud the picture. I don’t think it reflects on the vaccine,” Fauci told the newspaper.

While the circumstances of each of the company’s vaccine missteps has varied, a common thread has run through them all: a perception that the company has been less than fully transparent. The widening credibility chasm facing the AstraZeneca vaccine has severely eroded confidence in a vaccine that much of the world is still depending on to help end the pandemic.

Opaque numbers

At least once before the company has been caught potentially shading its vaccine data in a press release to make its inoculation seem more effective. On November 23, when AstraZeneca and the University of Oxford announced the preliminary results of its U.K. and Brazilian clinical trials, the press release trumpeted a 70.4% efficacy against symptomatic COVID-19. It did so again when a fuller analysis was published on December 8. But that figure was only obtained using a “blended” average that included a large group that received the planned dosing schedule of two full doses, where there was a 62.1% efficacy, and a much smaller group that had been given a half-dose initially, where the vaccine seemed to be 90% effective.

It turned out this smaller half-dose group consisted only of people younger than 55, meaning that besides being a smaller sample, there were other reasons to think that high efficacy number might not be representative of how the vaccine would perform in the general population. This was especially concerning because COVID-19 is most dangerous to older people.

It certainly looked as though the company had bent over backwards to provide a better headline number for efficacy—one that might put their vaccine at least in the same state, if not the same zip code, as the results announced a month earlier by Pfizer and Moderna for their vaccines. Both companies reported near 95% efficacy for their vaccines. Those two vaccines both use messenger RNA technology, while AstraZeneca’s vaccine is based on modifying a chimpanzee virus to carry instructions into the human body for making copies of the coronavirus’s spike protein, which then prompts an immune response.

The picture got murkier still as medical experts and reporters tried to get to the bottom of why some trial participants had been given the initial half-dose, with those behind the vaccine seemingly struggling to get their story straight. Mene Pangalos, AstraZeneca’s head of biopharmaceuticals research and development, told reporters that the half-dose had been a “useful mistake,” the result of a measuring error. But the company’s CEO, Pascal Soriot, told Bloomberg, “it was not a mistake.” And two Oxford scientists who developed the vaccine, Sarah Gilbert and Adrian Hill, both told journalists that there hadn’t been an error. Gilbert told The Financial Times that there “wasn’t a mix-up in dosing,” while Hill told Reuters, “that is really not true” the Oxford team was unaware of the half-doses when it administered them.

Finally, after a few weeks, a fuller story emerged: there had been a problem with the method Oxford—which was running the U.K. clinical trial—used to verify the amount of chimpanzee virus particles in the doses a contract manufacturer had produced for it. This method meant the test showed twice as many virus particles as should have been in each vial, even though the manufacturer, using a different method, said its own test showed the vials contained the correct amount. Because at the time Oxford couldn’t be certain which measurement was correct, and because it didn’t want to delay the trial, it got permission from the U.K. medical regulator to just give volunteers half doses from that particular batch of vaccine. Why AstraZeneca and Oxford couldn’t have explained all this to begin with is less clear.

And these weren’t even the only transparency issues surrounding the half-dose trial. Reuters reported that while Oxford had informed the doctors running the clinical trial about the half-dose, telling them it was a planned change in the study protocol, and the U.K. medical regulator was fully informed, it hadn’t told the trial participants themselves. "Presenting the dosing variation as a planned change in the study is potentially a breach of trust if in fact the dosing resulted from an error,” Emma Cave, a professor of healthcare law at Durham University's law school, told the news agency.

The dosing confusion may hurt AstraZeneca’s efforts to obtain U.S. approval for its vaccine. The muddled results increased the chances that the FDA would not accept the results of the clinical trials conducted outside the U.S. as a basis for emergency use authorization for the vaccine.

And those U.S. clinical trials were far behind schedule due to another problem that stemmed from AstraZeneca’s seeming tendency to be less than fully open in its initial communication. In early September, the company paused its clinical trials around the world due to a possible safety concern: one volunteer in its U.K. study had developed a serious neurological condition. But the company did not initially disclose the pause publicly, acknowledging it only after a reporter from the medical news site STAT News broke the story.

Worse, the company had not yet informed the FDA about the safety pause, meaning the regulator was blindsided by the STAT News report. Such pauses to investigate possible safety issues are routine in clinical trials, and they are not usually announced publicly. The FDA also gives those conducting trials seven days to inform it of possible serious safety incidents.

But the company should have realized that a “business as usual” communications strategy was not going to cut it. Given the pandemic, the entire world was fixated on the race to bring COVID-19 vaccines to market, and AstraZeneca had seemed like an early front-runner. What’s more, the world was particularly dependent on the company’s vaccine: because it can be stored at normal refrigerator temperatures and because it is inexpensive, the AstraZeneca vaccine is expected to play a key role in vaccinating people in low- and middle-income countries. At the time, it was also thought the vaccine would be used to inoculate about 60% of the U.S. population.

The company further compounded the uproar over whether it was being fully forthcoming, when AstraZeneca CEO Soriot later revealed to financial analysts that the trial had also been paused a few months earlier for another safety concern around similar neurological symptoms. In this case, the FDA had been informed. But the public had not. Nor had the company mentioned the previous pause in its initial public statements about the September safety stop.

While the independent review board looking into the neurological symptoms quickly concluded that there was no clear evidence linking the vaccine to the neurological symptoms the volunteer experienced and trials resumed in much of the world after just seven days, in the U.S. the pause stretched out to seven weeks. The regulator had requested additional data about previous vaccine trials Oxford had conducted using the same underlying technology. CNN reported that it took AstraZeneca an unusually long amount of time to give the FDA this additional information: more than a month. AstraZeneca says it struggled to get this data from Oxford in a format that was acceptable to the FDA. Whatever the reason, the long delay in restarting the U.S. trial did not help to boost confidence in the company’s vaccine.

Trouble in Europe

In fact, all of these problems have helped to undermine confidence in AstraZeneca’s vaccine. So much so that in many parts of Europe, a lot of people have been unwilling to take the vaccine, even when it is the only one on offer and large numbers of doses have gone unused. It hasn’t helped that European politicians and the German medical regulator openly questioned whether enough older adults had been included in the non-U.S. clinical trials to have confidence in the vaccine’s efficacy for those over the age of 65.

More recently, the company has had to defend the vaccine to medical authorities in a number of European countries who paused its rollout after a small number of cases in which people developed unusual blood clotting problems. The European Medical Authority has reiterated that the vaccine is safe: even if the clotting issue is linked to the vaccine—and so far there is no clear evidence that it is—it is an extremely rare side effect. The risk of severe COVID-19 is far worse.

The European regulators who decided to pause the rollout were almost certainly getting the balance-of-risks here wrong: the problem is, because of the previous stumbles in how AstraZeneca presented its results, confidence in the vaccine had already been severely undermined and fed suspicions about whether the company was being entirely truthful about the clotting issue too.

The company’s tendency to be economical with the truth has also plagued the commercial aspects of its vaccine rollout. The company has played up the fact that it is supplying the vaccine at cost “for the duration of the pandemic.” But The Financial Times reported that the company’s supply with various governments gave it the right to unilaterally declare the pandemic over as early as July 2021, a fact it had never mentioned publicly.

Similarly, the company drew the ire of European Union politicians after the company encountered manufacturing issues at the plant of a contractor in Belgium that was producing the vaccine. The problem meant AstraZeneca would be able to supply the EU with only a fraction of the vaccine doses it had promised. The problems were twofold: first, it took the company weeks to notify the EU. It has said it needed time to ascertain the nature of the problem in order to get an estimate for how big the impact would be. Secondly, Soriot infuriated EU officials by pointing out that its supply contract only said the company would make its “best effort” meet its delivery schedule.

Company officials have acknowledged privately that they’ve been slow to adapt their communication strategies—which might have worked well for normal drug development—for the fevered public scrutiny that has attended the coronavirus vaccines. They have also voiced frustration that the fact the company is producing its vaccine at no profit and has gone to extraordinary lengths to ensure it can play a central role in ensuring equitable vaccination access globally and yet it has the press or politicians outside the U.K. have cut it little slack.

Instead, these company insiders say, the company seems to be subjected to even closer scrutiny. Meanwhile, Pfizer has announced it will make $15 billion from its vaccine in 2021 alone and the public seems to shrug, these company executives grumble.

But then again, no one has accused Pfizer of trying to exaggerate its clinical trial data.