Rosie reincarnated
I couldn't help but think of Rosie as I read an article on the use of disinfecting robots at Johns Hopkins Hospital in today's Baltimore Sun. These robots seem to be all the rage. If you want to dig a little deeper, Clinical Infectious Diseases has the full account of the study done by Trish Perl's group. In the study, rooms that had previous occupants with MDROs were subjected to standard cleaning in 3 hospital units. In 3 other units, rooms were disinfected by hydrogen peroxide vapor producing robots. Subsequent room occupants were assessed for MDROs to determine whether transmission to the new occupant had occurred. The investigators demonstrated a significant reduction for VRE in subsequent occupants, but no significant decrease for MRSA, C. difficile, or MDR-gram negative rods. Of note, the study was partially funded via in-kind services by the robot maker. An excellent editorial by Cliff McDonald and Matt Arduino accompanies the paper.
The bottom line is that the authors demonstrated an absolute 6% reduction in VRE acquisition with use of robotic disinfection. Does that result warrant jumping on the robot band wagon? I don't think so for several reasons. First, with the exception of a few special patient populations (e.g., oncology and transplant patients), VRE has little pathogenic potential. Second, only a fraction of newly colonized VRE patients will develop infection. Third, a big problem with the study is that molecular typing was not performed. A few years ago, one of my IPs noted that 4 consecutive patients housed in the same room in our medical ICU all developed VRE bloodstream infections. I was certain that we had a problem with suboptimal cleaning between patients. Fortunately, we had all the isolates, and when molecular testing was performed we surprisingly found that the isolates were all genetically very distinct. It seems that VRE is an organism that while transmissible is also one for which antibiotic pressure is always creating new strains in individual patients. Lastly, even if this technology were perfect and rendered a room absolutely sterile, within seconds of the robot leaving and the humans returning, the room will once again be contaminated. It reminds me of the hysterical response by grade schools several years ago of shutting down a school and bleaching it after a child was found to have MRSA.
So to any hospital thinking about hiring Rosie, ponder long and hard, and consider taking all that money and using it to drive hand hygiene compliance to a new level. Remember, in health care, the hands remain the final common pathway.
Happy New Year!
Dear Mike,
ReplyDeleteThanks very much for your thoughtful and thought-provoking blog article about the CID article by Dr Perl’s group. My name is Jon Otter, one of the co-authors of the study. I’m the scientific director of Bioquell’s healthcare division, the “Rosie” under assessment in the study. I’m not a Jetsons fan, but I am rather fond of Wall-E and the cleaning robot comes to mind as a competitor to Rosie!
I’d like to pick up on a few points that you made in your blog post. Firstly, whilst the significant reduction was only achieved on VRE, the trend of reduced acquisition in the patients admitted to rooms disinfected using HPV was consistent for the other pathogens studied (MRSA, C. difficile and multidrug-resistant Gram-negatives). I appreciate that the whole point of statistics is to guide us as to what is due to chance and what is a genuine difference, but I am persuaded that this non-significant trend in the other pathogens is meaningful. In fact, this is why we performed the “combined” analysis by grouping together all four pathogens under assessment to conclude that patients admitted to rooms decontaminated using HPV were 64% less likely to acquire an MDRO when the prior room occupant was infected or colonised with an MDRO.
Secondly, it’s true that the absolute reduction in the incidence of VRE was in the region of 6%. However, remember that less than half of the rooms vacated by patients with MDROs were disinfected using HPV. If HPV had been used for the disinfection of every room vacated by a patient with an MDRO, extrapolating infection rates from all patient cohorts indicates that 17% of MDRO acquisitions would have been prevented.
Thirdly, I agree that nobody has “nailed” a prior room occupancy study demonstrating through molecular typing that patients can acquire pathogens from previous occupants of the same room. However, the evidence from observational epidemiological studies is overwhelming (reviewed by Otter et al. Infect Control Hosp Epidemiol 2011; 32: 687-699). Furthermore, there are now two intervention studies showing that you can mitigate the increased risk from the prior room occupant to a lesser or greater degree by improving the efficacy of discharge cleaning, which strengthens the epidemiological association further (the Hopkins study by Dr Perl’s group under discussion and this one from Dr Huang’s group: Datta et al. Arch Intern Med 2011; 171: 491-494.).
Finally, rooms will indeed quickly become recontaminated when patients shedding pathogens are admitted. This is why the most rational and evidence-based application of HPV and other automated ”no-touch” room disinfection systems is for the terminal disinfection of patient rooms, when the shedding patient has been discharged or transferred in order to protect the incoming patient (see recent review here: Otter et al. J Hosp Infect 2013; 83: 1-13). Recontamiantion caused by the incoming patient is not relevant when viewed from this perspective.
Of course, improving terminal room disinfection will only ever prevent a minority of transmissions. But I am persuaded by the prior room occupancy studies, meaning that we need to do a better job of terminal room disinfection whether achieved through improving conventional methods or adopting NTD systems in some circumstances, or, most likely, a combination of the two.