Discussion

In this study, the regional impact of intra-facility and inter-facility transmission of MRDOs/C. difficile among a tertiary care hospital and surrounding facilities was assessed in Japan. The unique healthcare setting in Japan was described in the patient population: the majority of the patients were the elderly; about 40% of the patients were admitted from surrounding LTCFs or hospitals; they were hospitalized for as long as a month; and more than a third of those originally admitted from home were not discharged back home. The molecular epidemiological analysis with POT method revealed that more than half of the target pathogens were detected early after admission, suggesting that MDROs/C. difficile were carried into the tertiary hospital from prehospital environments. In addition, the analysis confirmed that in-hospital contact (ie, intra-facility transmission) was a common pathway of transmission in the context of a relatively longer hospital length of stay at the tertiary care hospital. About two thirds of those discharged alive were discharged or transferred to LTCFs or other hospitals where the facility-level infection control was variable. The findings reinforce the knowledge of MDROs sharing network between LTCFs and hospitals. ^{Reference Gussin, McKinnell and Singh21,Reference Crnich22}

Using the POT method, we evaluated the possibility of MRSA/ESBL E. coli/C. difficile transmission in prehospital contact such as the residential setting at LTCFs before admission and its transmission during hospitalization at the tertiary care hospital. The POT method presents several strengths, including the ability to yield results within approximately half a day, whereas PFGE requires several days. It also facilitates standardized procedures using commercially available kits, and the numerical representation of POT types minimizes inter-examiner variability. However, a notable limitation is that the method is confined to detecting specific predetermined ORFs, which may diminish its capacity to differentiate novel strains not anticipated during its development. In phylogenetic analysis, the discriminatory ability of the POT method may be inferior to single nucleotide polymorphism (SNP) analysis using WGS ^{Reference Ogihara, Yamaguchi and Sato23} but equivalent to PFGE and MLST. ^{Reference Ogihara, Saito and Sawabe10} Another study showed Simpson’s diversity index for E. coli was 0.968 in the POT method and 0.979 in PFGE. ^{Reference Saeki, Sato and Furuya11}

The POT method has predominantly been applied to nosocomial outbreaks of MDROs such as MRSA ^{Reference Kawamura, Tokuda and Imuta24} and Acinetobacter baumannii. ^{Reference Fujikura, Hamamoto and Yuki25,Reference El Far, El-Mahallawy and Attia26} Others examined the concordance of C. difficile strains detected within a hospital ^{Reference Nakaie, Yamada and Kuwabara27} and explored strains of MRSA carried by companion animals. ^{Reference Taniguchi, Koide and Maeyama28} This study is unique in that the POT method was utilized to evaluate MDRO intra-facility transmission within a hospital during non-outbreak periods and to investigate the sharing of MDROs in the prehospital environment. Approximately one-third of in-hospital contact for MRSA were traced by the POT method, suggesting a potential burden of transmission within the tertiary care hospital. Conversely, the limited detection of prehospital MDROs sharing was partly attributed to the extensive number of facilities from which the patients were admitted (50 LTCFs and 20 hospitals).

POT1 for Staphylococcus aureus is composed of seven ORFs of which five are associated with SCCmec and two with the sequence type (ST). The combination of SCCmec and ST-type-dependent ORFs allows for the differentiation between conventional community-acquired MRSA (CA-MRSA) and hospital-acquired MRSA (HA-MRSA). Strains with SCCmec type IV, which predominantly represent CA strains, have a POT1 value of 104/106/108/110, while the New York/Japan clone (ST5, SCCmec type II), known for HA strains, exhibits a POT1 value of 93. ^{Reference Ogihara, Saito and Sawabe10,Reference Inomata, Yano and Tokuda12} Our research findings indicate that CA-MRSA strains were more frequently detected in specimens collected within 2 days of admission, while HA-MRSA strains were more prevalent in specimens collected after 3 days of admission. This trend is consistent with the conventional distinction between CA and HA strains. On the other hand, reports from Japan suggest that since 2010, there has been an increasing prevalence of SCCmec ²⁹ and POT strains ^{Reference Osaka, Okuzumi and Koide30} conventionally regarded as CA-MRSA detected in healthcare settings. Our findings of those with prehospital contact from the same LTCFs sharing the same POT codes, most commonly a CA-MRSA strain (POT 106-183-37), and those with in-hospital contact sharing the same POT codes, more than half of which was the CA-MRSA strain, suggest the epidemiological interaction may be more complex. In addition, the proportion of HA-MRSA (POT1 93) detected shortly after admission was 14.5% out of the total MRSA strains. This implies a potential burden of HA-MRSA in the prehospital environment, encompassing surrounding LTCFs and hospitals.

Among LTCFs where the assessment of HHSAF and IPCAF was conducted, the scores were generally low, Inadequate to Basic. We assessed two hospitals, one primary care and the other secondary care, and both yielded Intermediate to Advanced. A previous study across 57 hospitals in Japan reported that primary care hospitals tended to score lower in IPCAF compared to secondary or tertiary care hospitals, likely reflecting resource differences guided by the national regulatory system with financial incentives. ^{Reference Nomoto, Saito and Ishikane8} LTCF residents are frequently hospitalized, and even after discharge, they repeatedly get readmitted. ^{Reference Mor, Intrator and Feng31} Since LTCFs and primary care hospitals often serve as a source of referrals to tertiary care hospitals and as discharge destinations, improvement in facility-level infection control at these facilities may have a large impact on the regional AMR epidemiology.

This study has several limitations. First, we included only clinically collected cultures, and did not examine potential MDRO carriers who did not develop clinical symptoms. There is no standard policy for MDRO screening in healthcare settings in Japan, and the tertiary hospital of the study doesn’t perform routine MRDO screening. Active screening may facilitate a more comprehensive evaluation of MDRO transmission. Second, we didn’t collect microbiological specimens at LTCFs and other hospitals than the tertiary care, mainly due to resource restrictions and logistic feasibility. The molecular epidemiological analysis with the POT method among residents/patients at surrounding facilities may help us better understand the inter-facility MDRO transmission. Third, due to the retrospective nature of our study, prehospital and/or postdischarge healthcare data was not systemically obtained other than locations. Fourth, because this study was conducted during the COVID-19 pandemic when the visitation to facilities was restricted due to the health policy in Japan, the number of surrounding facilities that completed the HHSAF and IPCAF assessment was limited, precluding us from making more comprehensive assessment of the association between the epidemiological analysis and the infection control measures in a region, and even at the visited facilities, the situation of infection control might not reflect that in normal times. Fifth, this study was conducted at a single institution in Japan, which may limit generalizability to other regions and/or countries.

In conclusion, MDROs/C. difficile transmission is possibly prevalent within an acute care hospital and across its surrounding LTCFs and hospitals in Japan. The molecular epidemiology analysis with the POT method suggested intra-facility transmission may contribute to the regional AMR epidemiology in Japan where the care dependency is high in the aging society and the hospital length of stay is long at an acute care hospital. Further research is warranted on mechanisms of MDRO transmission at LTCFs, and how to prevent and control the inter-facility transmission with a region-wide approach.