Hospitals’ Infection Numbers Flawed

· Jan. 24, 2012, 3:55 pm

A new report says Colorado hospitals are getting a lot better at preventing one form of serious infections. But Colorado Public Radio Health Reporter Eric Whitney has taken a closer look and finds the numbers are fuzzy. 

CPR Health Reporter Eric Whitney: Every year the state health department publishes a report on how well Colorado hospitals do at preventing infections.

Sound: Paper report slapped onto table

Whitney: This year’s weighs in at 134 pages.

Sound: paging through report

Whitney: If you flip to the first page where results are posted, it looks like there’s some pretty good news. It says the number of bloodstream infections from central lines is down about 40% in the last four years. Central lines are tubes inserted in patients’ veins to deliver fluids or monitor blood.

A 40% drop in infections is a lot. But how reliable are the numbers?

It’s hard to say for sure, because the numbers the state publishes are what the hospitals report themselves. 

Crystal Berumen, Colorado Hospital Association VP for patient safety: They are, but they are validated by the health department. 

Whitney: Crystal Berumen is with the Colorado Hospital Association. She says the numbers are validated, but actually, they’re not.

I checked with Sarah Reese, who’s in charge of the report at the state health department.

It’s really impossible to say how accurate those numbers are?

Sara Reese, PhD., Patient Safety Program Coordinator, state health department: Correct, yes.  

Whitney: There’s two reasons the health department doesn’t check the infection rates hospitals report. One: Nobody says they have to, and two, the department doesn’t have the budget  to do it. It would cost hundreds of thousands of dollars.

Last year, though, for the first time, the health department was able to check up on a snapshot of the infection rates hospitals report. Sarah Reese: 

Reese: We found about 30% of cases were not reported, that should have been reported.

Whitney: Let’s be clear on that - the health department checked  the accuracy of one set of the numbers hospitals have to report. It found they were off by a third.

That’s not good, but the department isn’t saying all the infection rates hospitals report  are wrong, nobody’s ever checked the majority of them, so that’s impossible to say.

And those infection rates that are underreported? Reese says she doesn’t think hospitals are just doing it to make themselves look good. 

Reese: A lot of that came down to not understanding the surveillance definitions, there were some cases that were really complex, and they’re kind of referred to as microbial playgrounds, they just had multiple infections going on. 

Whitney: Reese says she thinks hospitals learned from having some of their numbers audited, and that they now more accurately report those types of infections.

But still, the state health department is using the hospitals’ flawed numbers to make a pretty concrete conclusion: That Colorado hospitals got 40% better at preventing certain bloodstream infections. That’s pointing to a four year trend when only three months worth of data has been verified. 

So how can you say there’s an overall decreasing trend if you don’t know the actual numbers from 3 years? Three and ¾ years I guess?

Reese: Well, you don’t, but if you don’t know the numbers for all – I guess in my opinion, if you’re consistently going down with the same consistent data you can’t boldly say that we’re going down, but you can recommend that we are going down. As I said, it’s not a perfect system, but it’s the best that we have.  So, you kind of want to be able to draw some conclusions from your data with the resources you have.

Whitney: So, even if the numbers aren’t 100% rock solid, the hospitals have been reporting, using the same methods across all four years

Reese: Right

Whitney: And so you think it’s legitimate to say it’s a 40% decrease because the data, even though it isn’t perfect, it’s consistent?

Reese: Exactly, yes.

Whitney: Other experts we checked with agree that, even though the bloodstream infection numbers aren’t 100% accurate, it’s still reasonable to conclude that there is a downward trend in how many are happening.

But what about infection rates for other procedures? No one knows how accurate they are, and Reese says some hospitals are better at reporting accurately than others. 

Reese: You have hospitals that have a lot of beds and they have huge surveillance programs, and a lot of IPs that are doing a lot of this work, and you have other hospitals that are just as big and they only have one IP…

Whitney: IP?

Reese: Sorry, infection preventionist, their surveillance is pretty basic, so they’re just bare bones, whereas you have other programs that are really gonna be going out on the floor, really looking for these infections.

WHITNEY: In other words, a hospital that’s good at finding its own infections will report higher numbers than the hospital that isn’t as diligent, meaning the good hospital will look bad, and the bad hospital will look good.

So as a tool for comparing hospitals, the state’s infection report isn’t very useful.

But, experts in the field say, that doesn’t mean it’s worthless.

Ned Calonge, former chief medical officer, Colorado Department of Public Health and Environment: (laughs) Two-thirds of the right answer is still so much better than none of the right answer. 

Whitney: Dr. Ned Calonge put out the first three hospital infection reports when he was the state health department’s chief medical officer. He’s

aware of its flaws, but says it’s evolving.

Calonge: as we get more accountability around the numbers, the reporting will get better. I think it can lead us down to more informed discussion, and I think that has value.

Whitney: More accountability and better numbers are coming. Colorado’s health department has won federal grants to audit reporting on four more infection types.

That should offer patients a clearer snapshot of how different hospitals compare at controlling them. 

[Photo: Centers for Disease Control and Prevention]

You care.

You want to know what is really going on these days, especially in Colorado. We can help you keep up.  The Lookout is a free, daily email newsletter with news and happenings from all over Colorado. Sign up here and we will see you in the morning!