The worst maritime oil spill in history began a year ago with a drop in pressure in a poorly drilled BP well deep in the Gulf of Mexico. (HO/AFP/GETTY IMAGES)

Still, troubles loom large for BP following its annus horribilis. It looks increasingly like a takeover target, reports say. Some shareholders are angry about high executive pay, while other investors remain worried about its safety practices or upset with the company for cutting the dividend.

But perhaps no challenge will be greater for BP than learning from the disaster and putting in place changes that keep such a catastrophe from ever happening again. After all, balance sheets can be bolstered with dollars, management structures can be reworked by moving people around, and claims funds that boost the local economy can be distributed by cutting a check. But human nature, that tricky root of any company culture, is not so easy to change.

BP’s chairman writes in the annual report that it is a “changed company” that has a “refocused strategy built on the pillars of safety, trust and value creation.” CEO Bob Dudley is reviewing how the company incentivizes and rewards people. He is reorganizing its business divisions, centralizing the company’s exploration efforts. And he has put a new executive in charge of safety, having him report directly to the CEO and embedding his staff within business units around the company.

In addition, writes Dudley, in order to “think hard about what was previously unthinkable,” BP is studying the nuclear and chemicals industries to see what it can learn from them. That’s the second time he uses the word “unthinkable” in the report: Just days before the Deepwater Horizon accident, Dudley writes, he had been “reflecting on the progress made by BP. The company had put safe and reliable operations at the centre of everything. …Then came the unthinkable.”

But was the explosion really so unimaginable? After all, BP had suffered through a refinery explosion in Texas that killed 15 people just five years before. And the number of signs of trouble leading up to the Deepwater Horizon explosion makes it hard to believe it was beyond the realm of comprehension. Crew members had already been calling it “the well from hell.” The National Oil Spill Commission’s report faults “a series of identifiable mistakes made by BP, Halliburton and Transocean that reveal such systematic failures in risk management that they place in doubt the safety culture of the entire industry.”

The problem, write business school professors Catherine H. Tinsley, Robin L. Dillon and Peter M. Madsen in the April issue of Harvard Business Review, is that the BP disaster was a “close to perfect” case study “in the anatomy of near misses and the consequences of misreading them.” Because so many wells on other rigs had suffered minor blowouts during cementing, the authors write, “the stakeholders were lulled into complacency”: “Each near miss, rather than raise alarms and prompt investigations, was taken as an indication that existing methods and safety procedures worked.”

The trio’s research finds that, in all of the disasters and business crises that it has studied, multiple near misses or close calls precede it. Managers become blinded by cognitive biases (anomalies start to be seen as normal over time) and outcome biases (past successes, despite near misses on the way, lead to a focus on results rather than processes). Because changing such biases “runs contrary to human nature,” they write, changing them, whether in an individual or an organization, is extraordinarily difficult.

Maybe by studying the nuclear and chemicals industries BP will pick up on the idea that avoiding catastrophe comes from learning from all the “near misses” that precede a disaster, and treating them as “instructive failures,” as Tinsley, Dillon and Madsen write, rather than as anomalies that could never happen here. But from the steps BP has outlined so far that it’s taking to learn from the Deepwater Horizon’s lessons, it’s not clear whether they will or not. Reorganize all the management structures you want, put in place all the safety functions you care to, rewrite all the incentive plans on the books, and you could still have an organization that doesn’t learn from—or make adjustments as a result of—the close calls and near misses that are inevitable harbingers of the disasters to come.

More from On Leadership:

FAA resignation: Accountability’s limits

Debt ceiling: When it’s time to break a promise