One likely issue is that it is hard to measure whether an error actually caused death. As the BMJ authors point out, medical error is not noted on the death certificate, nor does the coding system account for it. At best, medical errors have to be estimated using various imperfect methodologies. That is one reason why To Err is Human gave a range of between 44,000 and 98,000 deaths attributed to error. Another issue is the nature of death itself. Ultimately, death is not preventable, and many people come into a healthcare setting in compromised states. They then may get discharged into less than optimal care at home. Who is to say whether an error caused the death or whether it was the error combined with other factors?
Which made me think of baseball. In baseball an error is a judgment call, a determination made by someone knowledgeable about the game. While it certainly may contribute to the loss of the game, it must be weighed against other factors, such as the ability of the players and how they are feeling on that particular day. And no matter how good the team is, some games are going to be lost. There aren't undefeated seasons in baseball. Perhaps that is how we should think of medical errors. They are mistakes committed by professionals. The better the professional, the fewer there should be, but mistakes will happen. So perhaps we should focus, as we have been, on execution and on fundamentals that are tightly measured rather than on questionable big picture statistics. It helps to keep your eye on the ball.