As it turned out, almost all of the disruptions could be traced to the same root cause. In its effort to expand the reach of the EMR system, Partners had neglected to upgrade its operating system, and the old system just couldn't handle the load it was forced to bear. The IBM consultants confirmed this. The old system hadn't been designed to support a record system as large and dispersed as Partners'.

Upgrading the system, however, would take time, primarily because the vendor, Intersystems, had never worked with an application infrastructure as complex as Partners', and every piece of the new system would have to be tested. In the meantime, Glaser's team spent September making short-term fixes, such as adding more servers. By October, there were only four incidents, and by the end of December, the new upgraded system was in place. Glaser hopes the new operating system will keep the EMR database up 99.9 percent of the time. In the aftermath of the crisis, Glaser has had to reassess his plans to extend the EMR capabilities to all of Partners' affiliated doctors. That ambitious project will have to wait. "If you continue to roll out the implementation, you just look reckless, like you are so bent on the goal that you are ignoring reality," he says.

Instead, Glaser will focus on reestablishing his department's track record. "If you have a good reputation when you go into [a crisis], people believe that you are honest and that you can fix it," he says. "Your words have an aura of credibility. But if a problem goes on too long, history becomes irrelevant."

Glaser won't know for months the extent to which the prolonged disruptions have affected the credibility of his IT leadership. But he doesn't have time to worry about the repercussions now. He's focused on keeping the EMR system up and running 24/7, because that, after all, is the future of medicine.