Friday, June 13, 2008

Unconventional Wisdom

Subversive. Alternative. To put it as a cliché, "outside the box" thinking. That's what it sometimes takes to get the good ship US Medicine to change course. Over the past decade, a juggernaut composed of the US govermnent, large employers, insurers, and folks like the Leapfrog Group have been pushing a variety of changes and outright demands on physicians and hospitals. Many of these have some common sense foundations, and they all achieve the ever-important status of "feeling right," but most have never been widely tested.......this is simply following conventional wisdom. Nonetheless, we must comply (or else), with the US Medicine forced to follow the approved heading, which is may not always be the correct course.

One issue that has gotten little press but a lot of emphasis is the timing of prophylactic antibiotc administration for surgery in the prevention of surgical site infection (SSI). CMS, my good friends at JCAHO, and the CDC have basically outlined standards, following consensus guidelines, that prophylactic antibiotics should be given within 60 minutes before the incision to achieve therapeutic levels. This is the kind of guideline that nobody ever gripes too much about, because it makes logical sense --- give the prophylactic antibiotic on time, achieve therapeutic levels to achieve appropriate prophylaxis, and decrease your patient's risk of getting a wound infection. Sounds great!! Sign me up!!

One. Little. Problem. This hasn't ever been tested in the real world. That's kind of a big issue, because compliance with this guideline is going to be posted on the internet and will in the near future be tied to reimbursement. As well, the same regulatory agencies are now mandating which antibiotics are to be given, with somewhat conflicting data to stand on. Given the whole idea that Medicare and a variety of insurers are going to start refusing to pay for certain "never events," even though they do not describe how to prevent some of them (that's worth another month of posts), it is not a stretch to imagine the day when a surgical wound infection will fall into that category. Hence, a little data would be helpful.

I guess I wasn't the only one with this conundrum. Some folks from a variety of VA system hospitals actually looked at over 9,000 elective operations in 95 VA hospitals, with the premise that patients receiving timely prophylactic antibiotics should have lower rates of SSI. What they found was very interesting.

The study population included 9,195 elective procedures (5,981 orthopaedic, 1,966 colon, and 1,248 vascular) performed in 95 VA hospitals. Timely PA (prophylactic antibiotic administration) occurred in 86.4% of patients. Untimely PA was associated with a rate of SSI of 5.8%, compared with 4.6% in the timely group (odds ratio = 1.29, 95% CI 0.99, 1.67) in bivariable unadjusted analysis. Patient level risk-adjusted multivariable generalized estimating equation modeling found the SSI risk score was predictive of SSI (p < style="font-weight: bold;">SIP-1 (following the 60 minute guideline) was not associated with SSI. Hospital level multivariable generalized linear modeling found procedure mix (p < 0.001); SIP-1 was not associated with SSI. Hospital level multivariable generalized linear modeling found procedure mix (p < 0.0001), but not SIP-1 rate or facility volume, to be associated with facility SSI rate. The study had 80% power to detect a 1.75% difference for patient level SSI rates.

OK. Prophylactic antibiotics are good; that has been very well established with a large number of well designed clinical trials. But is there something magical about the time frame that we are being told to follow religiously or face public ridicule, loss of reimbursement, shame, scorn, and be forced to spend a weekend listening to Al Bore Gore? Well, no. As the authors of this study put it,
These data do not dispute the efficacy of timely PA administration for surgical procedures, but do call into question whether this process measure is the right metric for hospital quality of care for surgical patients. There are several explanations as to why we did not observe an association between timely antibiotic administration and SSI. The first is that timely antibiotic administration does not diminish SSI risk. This is an unlikely interpretation. There are numerous randomized controlled trials and observational studies that demonstrate the efficacy of prophylactic antibiotics in reducing SSI for various surgical procedures. Randomized controlled trials often have strict inclusion and exclusion criteria and a strict protocol that standardize the care with regard to management of the wound, and as a consequence, limit the external validity of the study. Our study included all patients undergoing indicated procedures for the quality measure. A more likely explanation is that the metric for timely antibiotic administration is too restrictive to be able to discriminate between prophylactic antibiotic practices that have a clinically meaningful effect on SSI prevention.Most patients in our study did get a prophylactic antibiotic; some just did not receive it in the timely window.
In the discussion that took place after this paper was presented, one of the authors made an interesting comment --- of those patients who did not receive their antibiotics in the "60 minute window," most (80%) were given too early. It does not matter whether that means 5 minutes too early to CMS or JCAHO; it's too early. But, as this study nicely points out, from a clinical standpoint, it doesn't matter (unless you are worried about getting paid) as long as the patient actually received the antibiotic.

Now, please don't get me wrong. I'm so stinking anal retentive that I never leave the OR to scrub after my patient is asleep until I have asked the anesthesiologist whether the patient has had their antibiotics (if needed), and checked to make sure they have received prophylactic subcutaneous Heparin or Lovenox (if needed), and checked to make sure they have SCDs on. That's at least what I feel is good medical care. But I do agree with these authors that if we are really going to hammer hospitals (and doctors by extension) for not following guidelines, it's pretty dang important to make sure that we have appropriate guidelines backed up by appropriate data.