CAUTI SCHMAUTI ! (part 2)
The November issue of Infection Control and Hospital Epidemiology has a retrospective cohort study by investigators at Baylor University. Over a 9-month period they examined all cases of positive urine cultures occurring in the presence of a urinary catheter in 5 medical wards and 5 extended care wards. This yielded 308 patients with catheter-associated bacteriuria. They went on to subclassify cases as CAUTIs and performed a 30-day follow-up. They found 2 cases of secondary bloodstream infection in the 128 CAUTI cases identified (1.6%). They cite a prior study by Dennis Maki that found 1 secondary BSI in 235 cases of CAUTI (0.4%). Now it's important to remember that the primary reason we are interested in CAUTI is the risk of secondary BSI because that's where we have the serious morbidity and mortality.
So let's to try to put this into perspective. We'll look at the worst case scenario first (i.e., we'll use the pre-intervention CUSP rate of infection [2.55 CAUTI/1,000 catheter days] and we'll use the Baylor risk of secondary BSI [1.6%]. If I apply these assumptions to my 850-bed hospital with 49,000 catheter days per year, I can expect 2 BSIs due to CAUTI yearly. Repeating this exercise with the CUSP post-intervention CAUTI rate [2.14/1,000 catheter days] and Maki's rate of secondary BSI [0.4%], I can expect 1 BSI due to CAUTI every 2 years. Now the 2012 rate of CAUTI at my hospital was actually 2.15 (essentially the same as the post-intervention rate in the CUSP report), so I can expect 1 BSI secondary to CAUTI every 7 months to 2 years. When I think about the resources and energy that we are expending to reduce our CAUTI rate, I have to conclude that there's not much bang for the buck. Put more eloquently, the opportunity cost is high. Or perhaps more to the point: chasing CAUTI is a fool's errand.
Photo: Saltanat Ebli, Wikimedia
With 34,000,000 discharges a year of patients 18+, what would be a number of CAUTI cases that you think would be ok? How about 340,000? For it to happen to 1% of patients, would that be OK? If 340k isn't about right, would a good goal be a higher or lower number?
ReplyDeleteHere's another way to look at it: if you assume that 1% of CAUTIs lead to a BSI, then the impact of 1 CLABSI is equivalent to 100 CAUTIs. So if a hospital has not driven their CLABSI rate to the irreducible minimum, where should they direct their resources?
ReplyDeleteAre CAUTIs that don't lead to BSIs inconsequential? They may be. Is that your point? That 99% of CAUTIs don't matter because they don't lead to BSIs. I don't think I've heard that point of view stated so clearly before. If that's what you're saying.
ReplyDeleteThat is my point. 99% of CAUTIs are low impact infections. They may cause some discomfort (although many are asymptomatic), but don't add to length of stay. Accordingly, the cost is relatively low, in comparison to CLABSI and VAP. Most hospital infection programs still have plenty of work to do on the high impact infections, as well as hand hygiene.
ReplyDeleteShould we ignore indoor radon until we convince all smokers not to smoke? Should we not install air bags in cars until we get 100% compliance on seatbelts?
ReplyDeleteWhy do you assume it's a zero sum game? People running hospitals could be having personnel working hard to bring their CLABSI rates down AND also have people working to get the CAUTI rates down at the same time. For one thing, it seems like it might be different people in some cases, the latter could focus in part on surgeons telling nurses to take out urinary catheters ASAP, how does that get in the way of CLABSI reduction programs that mostly involve the insertion of central lines by a completely different cadre of professionals? Maybe the ICP and Hospital Epidemiologist have to pay attention to both efforts, but that's a small part of the job compared to what the people inserting and removing lines have to do to bring down CLABSI and CAUTI rates. At least that's how it seems to me. (And I've had a couple of urinary catheters put in and removed in the last year, and one central line in and out. None were infected. My surgical site? That's another story. But it has had a happy ending overall.) Am I missing something, is it too hard for the ICP and Hospital Epidemiologist to think about two reduction programs that begin with C and end with I at the same time? When people ask how did we get in this situation in healthcare, this is a good example. If we have to wait until CLABSI is conquered to work on CAUTI, then fix that before starting on SSIs, then fix that before working on pressure ulcers, then fix that before we start on medication related adverse events... in the meantime all these things add up annually to literally millions of adverse events or hospital acquired conditions.
As someone on the frontline I can tell you that the scope of work for Infection Prevention programs continues to expand, and in most places the resources haven't kept up. I don't know of any program with unlimited resources, so it really is a zero-sum game for the most part. I have been fortunate to have a program with good resources, but many IP programs are struggling. All programs have to prioritize their work and my point is that CAUTI should be given a low priority given its low clinical impact and relatively few interventions shown to be effective in rate reduction. The goal of any good infection prevention program should be first and foremost to reduce mortality, and until and unless the hospital has driven their rates of HAIs with high attributable mortality down to close to the irreducible minimum, they are foolish to put much energy into CAUTI. My goal as a hospital epidemiologist is clearly utilitarian. I want to do the greatest good. Chasing CAUTI won't get me there.
ReplyDeleteShould C diff and SSI have low priority or just CAUTI?
ReplyDeleteCLABSI has a higher mortality rate than both of these too. Or should priorities be based on total deaths or death rates? If infection A has twice the death rate of the infection B but causes half the deaths, because A is much less frequent is A the more important one or the less important one? Is it utilitarian to focus on B because it kills more people? How about if one doesn't kill frequently but sometimes causes serious non-fatal outcomes like colectomies, how does that figure in? Do 10 colectomies equal one death? Does the age of the typical patient matter? Or is it just that you know that CAUTIs don't matter because they don't kill that many people, and because they can easily be treated with antibiotics. Unless all the antibiotics are starting to not work which seems to be a recurring theme too... Maybe just let someone else in your hospital work on CAUTI reduction and you can focus on the important stuff. Or don't, and tell all you non-physician friends that if they come to your hospital they shouldn't be surprised if they get a post-op CAUTI because you don't think it's worth anyone's time to implement simple interventions like asking the surgeon every day if it's OK yet to take this catheter out, or even taking it out before the patient wakes up from the operation if it's pretty clear that the patient will be ambulatory and meets some other simple criteria.
Went back and read your first entry about rates again. Does 49,000 line-days and rate of 2.55 equal 125 CAUTIs? On-line it looked like your hospital had 56,000 discharges in 2012. Does this mean you have 1 CAUTI per about 450 discharges? That does sound low! Do you think the CAUTI definition (assuming NHSN) is too restrictive? i.e., that many cases of what may be CAUTIs, or are at least treated as CAUTIs with antibiotics, are not being counted? Or do you think 125 treated CAUTIs is the same as the official number? It seems that national estimates of CAUTIs have gone down a lot in recent years, even as the SIR has not gone down.
ReplyDeleteThe 56,000 number you cite is adjusted discharges (I'm not sure exactly what that is). The actual inpatient discharge number is 33,000, so infections per discharge is higher than your calculation. I don't think the NHSN CAUTI definition is too restrictive; if anything it overcalls UTIs (in the last quarter 1/3 of our CAUTIs were due to Candida, some of which were probably were not truly UTIs).
ReplyDeleteOk. Thanks for the correction. Let's say you have 33,000 discharges and the country has 33,000,000 (pretty close). Let's say your hospital has 100 real CAUTIs (-20%) rather than the 125 estimated above. If the rest of the country does the same as your hospital that would be 100,000 real CAUTIs nationwide. This would be one CAUTI per 330 discharges, not 1 per 450 as estimated above. The strange thing is that in 2009 CDC estimated the national count of CAUTIs in inpatients as 449,334, about 1 in 75 (see http://www.cdc.gov/hai/pdfs/hai/scott_costpaper.pdf and http://www.cdc.gov/hai/pdfs/hai/infections_deaths.pdf). This info was for 2002. So what do you think is going on? Some options... Has more than 80% of the CAUTIs gone away in the last decade (seems unlikely)? Is your hospital 4 times better than the rest of the country? Has the CAUTI definition changed? Was the 2002/2009 CDC number totally wrong? Is your 2.55/1,000 number wrong (for example, correct according to the current definition minus Candida, but wrong according to what CDC was calling a CAUTI in 2002)? Something else?
ReplyDeleteSeparate question: do you think that you are really only giving 100 people per year (2 per week) antibiotic treatment for hospital acquired CAUTIs in your hospital? It seems that at some other hospitals (and in other settings like urgent care, etc.) people may be getting treated for presumed CAUTIs - in the absence of a positive test or prior to a positive test - based on symptoms. What do you think of the idea of counting CAUTIs based on doctors diagnosing and treating them like CAUTIs, even if they are not always confirmed as CAUTIs?
The NHSN definition has changed and may account for some of the discrepancy. Asymptomatic bacteriruria used to be included and that was dropped I think in 2009. I am not in favor of basing the definition on doctor diagnosis/treatment as it would introduce a huge amount of variability.
ReplyDeleteThis comment has been removed by the author.
ReplyDelete(Removed comment above and fixed typo.) Looking at your estimates and my estimates above it comes out to about 1,600 BSIs from CAUTI a year (1.6% and 100,000 CAUTIs: using the NHSN definition and extrapolating up - my guess at a national number based on the rate at your hospital). Or about one per three US hospitals. Do you think this is a credible estimate? Do you get only about 1 or 2 post-CAUTI BSIs a year at your (very large) hospital?
ReplyDeleteHow about this paper that suggests that 1.1% of all hospitalized patients get catheter related UTIs? That would be close to 400,000 for the US. Would this still be "schmauti"? The first author was just elected to the IOM this year so its unliklely that he's a total crackpot... http://qualitysafety.bmj.com/content/22/10/809.full
ReplyDelete