The Silent Killer: Hospital-Acquired Infections

From the archives, comes this post about the health crisis no one cares about (except for the Mad Biologist. We are very caring): the 90,000 deaths per year from infections people get while in the hospital. And this number is probably an underestimate.

Bacterial infections aren't sexy: no one walks, bikes, hops, pogo sticks for the cure. There are no ribbons, no bumper stickers, and no hot celebrities (damn!). Yet, according to the CDC, bacterial infections acquired in hospitals kill at least 90,000 people per year in the U.S. Granted some of those who died would have died from something else anyway, but that's still a really large number. To put this figure in context, in 2004, roughly 18,000 people died from HIV. 40,000 died from breast cancer.

Today's Washington Post discusses the data from Pennsylvania:

12,000 Pennsylvanians contracted infections during a hospital stay in 2004, costing an extra $2 billion in care and at least 1,500 preventable deaths...

Here's the awful part: this number is probably a ridiculous underestimate. Health insurance claims indicate that the number of hospital-acquired infections in PA was around 115,000. It's probably not that high either: some infections may have been acquired outside of the hospital, and some hospitals may be gouging insurance companies.

However, the true number is probably pretty close to 115,000. First, the researchers removed many cases that might have been acquired before entering the hospital. Second, you charge insurance companies for $5 Tylenol and unneccessary consultations. You don't pump people full of amoxicillin just for the hell of it (Intelligent Designer, I hope not). So the 115,000 figure is probably pretty close.

So what kinds of infections are occurring in hospitals? The number one type of infection is urinary tract infections--all those wonderful catheters that poke through natural biological barriers that serve as the first line of defense against infection. (I'm not starting an anti-catheter movement here; you gotta pee. This, however, is just the way the biology works). The most lethal infections are respiratory infections (pneumonia) which kill 32% of patients who get these infections. These are usually associated with respirators.

Here's a pretty chart that breaks it down for ya:


Remember the quote from the top:

12,000 Pennsylvanians contracted infections during a hospital stay in 2004, costing an extra $2 billion in care and at least 1,500 preventable deaths...

Keep in mind that this could be off by a factor of ten. PA is 4% of the U.S. population, so now multiply everything by twenty-five. This kinda puts that whole shark attack thing in perspective, doesn't it?

I was going to end with the snarky comment above, but the more I think about this, the angrier I get. Look, the doings of Turd Blossom are important, no doubt about it. But this crisis--and 90,000 dead constitutes a crisis--is just as important. As the article notes, this isn't a new problem either. But it's not 'sexy.' It's a 'science/health' story, so it's not as important as the manly stuff. You have to know a little biology to understand the problem. You have to be numerically literate. And shock jocks and pundits don't bloviate about this because they're too ignorant of medicine and biology, so there's no hot air propelling this ship.

CDC estimates: 90,000 dead. 2,000,000 infected. Maybe if we called this an epidemic or a bacterial insurgency, people would pay attention.

Friends have often asked me why I don't take bioterrorism very seriously. I have a lot of reasons, but here's the germane one: the bioterrorist attack, for all intensive purposes, is already here. 90,000 dead per year--forget the World Trade Center, that's a Nagasaki sized number.

Meanwhile, antibiotic resistance is on the rise, so don't think the chance of dying from an infection is going to decrease. If anything it will increase. And don't think this doesn't affect you. At some point, many people will end up on respirators or catheters, and not just the very elderly. Twenty years from now, you don't want to ask the doctor what to do, only to have her say, "Morphine. Because the antibiotics won't beat the staph infection."

The Mad Biologist is very mad about this...

More like this

One of the most interesting comments I saw on this issue was from someone whose job was to design buildings for the food industry. She was shocked when she went to hospital as she could see that the hospital was designed to be very hard to clean and by her reckoning basically impossible to keep clean enough to reduce bacterial levels. She stated that if you wanted to keep a building to low levels of bacterial contamination, then it has to be designed from the start to minimise places for bacterial growth and maximise ease of cleaning. She stated that no hospital she had visited would come close to being able to meet the standards of the food industry and none would be accredited as a food manufacturing site.

By Anonymous (not verified) on 29 Jul 2006 #permalink

We had a friend who came through the heart transplant just fine -- then died of encephalitis while still in the hospital. Any numbers on how many of those infections were in patients who took immunosupressants??


My condolences. There's virtually no data on the role of immunosupressants (I also don't know if the encephalitis was bacterial or viral). Most of the mortality and morbidity data are based on hospital discharge codes. Often, these codes are mysteriously not entered when there is a hospital-acquired infection, particularly when there are other causes. Things like immunosupressants are even harder since these aren't coded, requiring a case-by-case review (which is difficult unless the government mandates a record release, or all patients release the records).

The short answer: I have no idea.

I second your latter comment Mike, and can you bet that the "accidnetal" omission of nosocomial infection codes will become more of a standard once we get into the swing of mandatory quality data reporting on hospitals. Just like in the early days of HIV positive samples, it will be the laboratorians who will be able to know the state of the hospitals' health.

Sorry for your loss Scorpio, it is another sad reminder that hospitals are NOT places for sick people.

By Chromosome Crawl (not verified) on 30 Jul 2006 #permalink