Robots Have Killed 114 People Over a 14-year Period, But We’re Not Worried

The BBC had a report on this study that said 144 people died over a 14-year period in incidents involving robotic equipment during surgery a couple days ago, part of what seems to be a media trend of finding fertile ground in fear of our soon-to-be robot overlords. What they didn’t talk about much is that this is more good news than bad — and that the bad news involves humans more than robots.

That sounds callous — 144 people died, and that’s 144 tragedies. If medical technology can continue to improve to cut that number down even further, that’s great. The good news is that this is an average of about 10 deaths per year, coupled with the finding that deaths per year peaked in 2007 (important to note that 33.7 percent of incidents of death were due to risks inherent to surgery). Granted, the study warns of underreporting — they’re using the FDA MAUDE database’s reports on robotic surgical systems, and no database is perfect — but even then, the number is impressively low for what is still a relatively young technology. It seems like a rare case of a tech sector getting things right the first time — as well it should be, with lives at stake.

The more alarming part of the study is that there are so many gaps in it, through no fault of the researchers. The study is plagued with incidents reports that failed to note which kind of surgery was being administered during the incident. Another big asterisk in the study is that the researchers couldn’t directly report rates of death and injury per procedure using robotic equipment, because the FDA’s database warns that rules for reporting have been in flux and that its data are not reliable enough to establish those rates. Any rates given in the study are estimates from the researchers.

The numbers from the report themselves are promising — from the researchers’ estimates, we still have far more to fear from being in a hospital in the first place — healthcare-acquired infections still affect over one million people per year, with deaths in the tens of thousands. That can’t directly be compared to the robotic surgery numbers, because people undergoing procedures involving robotic equipment is a relatively small subset of total number of people who visit a hospital in one year, but the difference is stark enough, and the study’s estimates of rates of death and injury (given for each kind of procedure) indicate that they are highly unlikely.

The technology isn’t what needs improving at least, it shouldn’t be the number one priority. As we start to lean on robotic equipment more and more, consistent standards and regulations related to incident reporting are necessary to properly understanding what needs fixing in the first place. It’s good to have studies like this, but without quality data, pinpointing the most pressing dangers related to robotic surgery seems impossible.