September 2019: Antimicrobial stewardship and infection control

What determines which healthcare personnel actually accept and complete treatment for latent TB infection?  The authors retrospectively reviewed the occupational health records at their large academic medical center over a 10-year period, identifying the employees who had positive tests for latent TB infection upon hire and examining the factors associated with accepting and completing LTBI treatment.  The center required employees with a positive LTBI screen to attend an employee health clinic visit to discuss LTBI treatment, but did not require LTBI treatment initiation or completion as requirements for continued employment.

A total 570 employees were found to have latent TB infection, and 470 of these were offered treatment; reasons to not be offered treatment included prior treatment for LTBI (n=39), failure to attend the clinic visit (n=34), and contraindication or comorbidity precluding treatment (n=27).   The majority of employees were male, the median age was 35, and the most common countries of origin were India, China, Korea, and the US.  Of the 470 offered treatment, just 193 employees accepted; 173 were given 9 months of INH and 20 were given 4 months of rifampin. Just 137 employees (29%) completed treatment- 118/173 employees offered INH and 19/20 offered rifampin.

In multivariate logistic regression, acceptance of LTBI treatment was associated with younger age (OR 0.96; 95% CI 0.94-0.99), male gender (OR 1.8; 95% CI 1.2-2.8), and employment as a physician (OR 0.5; 95% CI 0.3-0.8) or researcher (OR 0.4; 95% CI 0.2-0.7).  For the record, the ORs for nurses and students were about as bad, though they did not reach statistical significance at the traditional p=0.05 threshold.  Factors associated with treatment completion included coming from a medium or low TB-incidence rather than a high-TB incidence country (OR 8.5 with 95% CI 3.9-18.4 and OR 9.5 with 95% CI 2.1-43.8, respectively), and having had BCG vaccination (OR 0.1 with 95% CI 0.01-1).  In addition, those who received 4 months of rifampin were much more likely to complete treatment than those given 9 months of INH (95% vs 68%; p=0.005), a finding that has been reproduced elsewhere.

Long story short, older folks, women, physicains and researchers are less likely to sign up for LTBI treatment, and folks who come from high-endemicity areas and have had the BCG vaccine are less likely to complete treatment if they sign up for it.  But the one actionable item here is that use of the 4mo rifampin regimen is associated with a fantastically higher rate of cure.  One wonders: if 1mo INH-rifapentine was offered for LTBI, as in the BREIF-TB trial, would completion be ever higher – and more importantly, would more employees agree to undergo treatment? 31552416

Which parts of a patient’s room do healthcare workers most often touch?  For two years, the authors had a trained observer watch adults hospitalized with acute viral respiratory infections over 3-hour intervals, recording each healthcare worker who entered the patient’s room, what kind of care they were providing, what surfaces they touched, and how often they touched them.  The investigators categorized healthcare workers as providers (attending and resident physicians, NPs, and medical students, n=41), nurses and nursing students (n=69), nurse technicians (n=37), respiratory therapists (n=9), and others (n=8 including physical therapists, environmental service workers, and dieticians).

A total 52 patients were observed; of these, 30 were on droplet + contact isolation, 21 droplet isolation only, and 1 contact isolation only.  A total 107 healthcare workers were observed having a total 166 interactions.  HCWs most often touched patients, and touched them a median twice per visit and 11 times per hour.  The most commonly touched fomites included the tray table (48% of encounters), bed surface (41%), bed rail (41%), computer station (37%), and IV pole (32%).  Healthcare workers touched themselves in 63% of encounters, most often touching their torsos (32%), masks (29%), lower bodies (28%), and hands (27%), for a median 1 self-contact per interaction and 5 self-contacts per hour.  Providers made significantly fewer fomite contacts than nurses and nurse technicians (median 3 vs 9-10 contacts per visit; p<0.05).

These data may seem banal, but they ought to inform where the environment services team spends their energy cleaning rooms during admissions and after discharge.  That is to say, focusing on keeping just a few high touch surfaces clean (the tray table, the bed and rail, the computer, and the IV pole) might pay big dividends in reducing the transmission of nococomial pathogens in your hospital. 31517975

A pharmacist-led prospective audit and feedback strategy effectively reduced the rate of antianaerobic combination therapy.  The authors, out of South Korea, initiated a pharmacist-led program to reduce the rate of combination antianaerobic therapy at their large (>1,000 bed) hospital.  They targeted combinations of either clindamycin or metronidazole with cephamycins, beta-lactam/beta-lactamase inhibitor combos, carbapenems, moxifloxacin, or tigecycline, and excluded cases in which clindamycin or metronidazole were added to treat C.difficile infection, toxic shock syndrome, or PJP pneumonia, or if the combination was being administered at the direction of the infectious disease team.  When a redundant combination was identified, the PharmD entered a template note in the patient’s chart explaining the redundancy and making alternate recommendations for antimicrobial therapy; secondarily, an ID physician reviewed and cosigned the note.  No follow-up notes were entered, nor were the primary physicians contacted or incentivized to change their prescribing behavior in any way.

The authors collected data on antianaerobic prescribing for 18 months before and after implementing the intervention, and they identified 737 and 768 instances of combination prescribing in the before and after periods, respectively. In the preintervention period, 50% of these combinations were classified as redundant, whereas only 35% were deemed redundant after the intervention.  In both periods, concurrent treatment for C.difficile was the most common acceptable reason for use of two drugs with anaerobic activity, and combination of metronidazole or clindamycin with a cephamycin was the most common redundant combination (51% of all cases of redundant therapy).  After the intervention, the median monthly incidence of redundant antianaerobic prescribing decreased from 5.3 per 1000 admissions to 3.3 per 1000 admissions (p<0.001); total use of clindamycin and metronidazole also decreased, from 3.3 days of therapy per 1000 patient-days to 1.7 days of therapy per 1000 patient-days (p<0.001).  Importantly, acceptable use of double anaerobic coverage did not change in incidence after the intervention.  The workload associated with this intervention also decreased over time, as improved prescribing behavior among the physicians led to fewer signed recommendations per 1000 admissions, down to less than 0.1 notes / 1000 admissions by 2018. 31482420

Cognitive biases regarding asymptomatic bacteriuria are widespread in the hospital setting.  That’s not actually the main point of this paper, but it’s what I want to focus on.  Drekonja et al performed a survey of attending and resident physicians, advanced practice providers (NPs and PAs), nurses, and clinical nurse assistants at four VA hospitals.  The survey contained questions assessing knowledge about the diagnosis and treatment of ASB and CAUTI based on practice guidelines as well as questions about behavioral constructs (i.e. social norms about practicing in real life & assessment of self-efficacy / perceived autonomy); the surveys varied in content between the CNAs, nurses, and other clinicians, but I don’t want to get into the weeds discussing this.  This was all part of a roll-out of a new antibiotic & diagnostic stewardship campaign for UTI, the “Kicking CAUTI” program, which is the brainchild of the senior author and my mentor Dr. Barbara Trautner (that’s my primary disclosure – but also I know and/or am working with some of the other authors at present).

Anyway, the authors collected 534 surveys: 48% from ‘providers’ (I hate this term and hate it especially when it’s used to mean doctors but not nurses), 37% from nurses, and 17% from CNAs.  The provider responses came mostly from residents (66%) followed by attendings (23%), and their overall scores on the knowledge questions were a middling 78%.  In particular, providers desire to treat asymptomatic bacteriuria varied by the organism isolated, with just 8% recommending treatment of ASB with mixed gram-positive flora but 42% recommending treatment when the organism was E.coli and 53% recommending treatment when the organism was an ESBL E.coli.  Ratings of self-efficacy and behavior constructs were low, indicating that while providers expressed confidence in the quality of UTI and ASB clinical guidelines, they felt that cultural and structural barriers kept them from adhering to the guidelines in practice.  Interestingly, providers who reported working on more functional teams (that is to say, they gave their teams high marks for teamwork and a culture of safety) were more likely to recommend appropriate management for ASB.

The nurses’ survey focused on appropriate reasons to get a urine culture, and here misperceptions were common, with 79% suggesting cloudy urine indicated UTI, 86% putting faith in foul smelling urine, and 49% indicating that change in urine color heralded UTI (none of these are true).  In addition, 51% believed that patients with catheters ought to have screening urine cultures (no), and 69% believed that pyuria distinguishes ASB from UTI (also no).  The CNAs were asked similar questions and gave similar results.

The take home here is that misconceptions about what defines ASB and UTI and how they ought to be diagnosed and managed are rampant among physicians, APPs, nurses, and nurse assistants.  Perhaps more importantly, though, inappropriate practice behaviors seem to be driven as much by cultural factors in the workplace as by simple deficits in knowledge. 31339085

N95 respirators were no better than medical masks for preventing influenza among healthcare workers.  No big surprise given that the ocular mucosa are probably also a route of entry for influenza, meaning neither mask is great protection on its own.  This study used a randomized cluster design and involved 137 outpatient centers over four 12-week peak influenza seasons.  There were 400 lab-confirmed episodes of influenza among 5180 person-seasons, with no difference between use of N95 respirators vs medical masks and in fact possibly lower rates with the medical masks (aOR 1.2; 95% CI 1.0-1.5) despite similar rates of mask adherence between groups.  The accompanying editorial points out the study was underpowered for its prespecified outcome; I’m not losing too much sleep over this point, though, given that we have a trial with over a thousand subjects in each arm yielding almost the exact same rates of influenza.  Without more compelling data for N95s, I think their cost and burdensomeness ought to preclude their routine use for influenza transmission prevention.  Disclaimer: one of the co-authors, Dr. Rodriguez-Barradas, is one of my mentors from fellowship.  31479137

Surgical site infection risk increases with BMI category from normal (18.5-25) to morbidly obese (>40), and this is true for a wide variety of surgeries.  These conclusions come from a retrospective review of over 387,000 patients who underwent procedures ranging from CABG to mastectomy to colectomy to knee arthroplasty to caesarean section.  The most profound association occurred with hip arthroplasty, where morbid obesity was associated with 7.8-fold risk for SSI versus normal weight; the weakest association was with hysterectomy, and this case  being overweight (BMI 25-30) actually appeared protective (RR 0.3 with 95% CI 0.1-1.0), and no association at all was found with laminectomy.  Important confounders in these analyses included the patient’s age, gender, and the duration of the surgery in question.  31232239