Discussion

In our survey of US SRN acute-care hospitals, most facilities performed prospective HA-IMI surveillance (69%) and most utilized air sampling for mold (62%) as part of HA-IMI prevention or investigation efforts. However, both HA-IMI case definitions and approaches to environmental sampling for mold varied substantially among facilities. The relatively high percentage of participants engaged in HA-IMI surveillance likely reflects that respondents were mostly from academic institutions caring for patient populations at high risk for IMIs (eg, patients with hematologic malignancies or receiving stem cell transplants). This finding is consistent with guidelines from the Infectious Diseases Society of America, which strongly recommend that leukemia and transplant centers surveil for cases of invasive mold infection. ^{Reference Patterson, Thompson and Denning7}

Facilities differed in both how IMIs were defined and how they were determined to be healthcare-associated, mirroring the diversity of case definitions that have been applied in previous HA-IMI cluster investigations. ^{Reference Hartnett, Jackson and Perkins3} Ascertaining whether IMI cases are healthcare-associated is difficult because patients who develop IMIs often have complicated medical histories with multiple possible exposures and the incubation periods for some mold infections are not well established. ^{Reference Hartnett, Jackson and Perkins3} Furthermore, certain hospitals may lack the laboratory capacity for prompt mold species identification. Despite these challenges, systematic surveillance for HA-IMIs is a necessary step in understanding disease burden, quickly identifying potential clusters, and reducing mortality from these infections; therefore, efforts to develop a feasible HA-IMI surveillance approach, including a standardized HA-IMI case definition, should be prioritized.

Hospitals varied substantially regarding whether and how they performed air sampling for mold, reflecting the controversial role of air sampling in preventing HA-IMIs. During previous HA-IMI clusters, air sampling has been helpful in identifying potential targets for remediation or supporting possible epidemiologic links between a mold sources and affected patients. ^{Reference Hartnett, Jackson and Perkins3} For such investigations, air sampling is most useful when performed as an adjunct to a detailed environmental assessment and epidemiologic investigation using a well-designed sampling plan, optimally under the guidance of an industrial hygienist with experience participating in microbiological assessments. ^{Reference Hartnett, Jackson and Perkins3} Air sampling may also be useful for monitoring mold counts before, during, and after major construction activities, a practice recommended by guidelines in several countries, ^{Reference Talento, Fitzgerald, Redington, O’Sullivan, Fenelon and Rogers8,9} but it is not currently recommended by the US Healthcare Infection Control Practices Advisory Committee. ^{Reference Sehulster and Chinn10} The utility of environmental air sampling may be limited by the lack of established threshold values or regulatory levels for mold in air and the lack of widely accepted industry qualification or practice standards for mold assessors and remediators. ^{Reference Hung, Caufield and Miller5} Efforts to identify and close knowledge gaps regarding air sampling strategies for mold are needed; in the interim, consensus guidelines based on expert opinion and existing literature might empower acute care facilities to adopt rational approaches to air sampling for mold.

Our findings are limited by our inability to follow up with respondents and our small sample size. The HA-IMI surveillance and air sampling practices reported by the facilities surveyed might not represent practices of acute care hospitals nationwide. We suspect that facilities with interest or prior experience in HA-IMI prevention were more likely to respond to our survey and that HA-IMI surveillance and air sampling for mold may be less common in other hospitals. Nonetheless, our findings underscore the need to develop generalizable strategies for HA-IMI surveillance and for further data to guide rational approaches to air sampling for mold.