Predictors of Growth of Escherichia Coli on Lab Coats as Part of Hospital-Acquired Infection Transmission Through Healthcare Personnel Attire

Background: Healthcare personnel (HCP) attire could be potential vehicles for microbial transmission. This study aimed to explore the extent of bacterial contamination and hygiene and handling practices of healthcare personnel attire that could inuence bacterial growth. Methods: Descriptive, cross-sectional study was used in this study. Convenience sampling of the 188 healthcare personnel was recruited from a main holistic hospital in the northern part of Jordan. Three swab samples were collected from three different parts of lab coats used by each participant. ANOVA test and the generalized mixed linear model were used for the categorical variables of three or more levels and identify the predictors of positive growth of Escherichia Coli on healthcare personnel attire. Results: Enterococcus faecalis was the most common species of bacteria found on lab coat. Despite of no statistically signicant differences were found, the HCP attire coming from the critical care units and the emergency department were highlighted with slightly higher contamination compared to other departments. Escherichia coli ‘s near-signicant growth differences (p=0.057) were found based on different locations on lab coats compared to other types of bacteria. Factors associated with signicant growth of Escherichia coli on healthcare personnel attire were age ≥ 36 years, a high income, bachelor marital status, working as a physician, lab coat use preferred over scrubs, and borrowing of lab coats. Conclusion: Healthcare personnel should be cautious about the method of use and storage of lab coats they wear.

and cause severe illnesses and deaths in hospitalized patients, are clustered in what is known as the ESKAPE group (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species). The hazard of this group is due to their multidrug resistance. 17 The species in the ESKAPE group that were found to be transmitted through lab coats include a methicillin-resistant Staphylococcus spp., Pseudomonas spp., Klebsiella spp., Escherichia coli (E coli), and vancomycin-resistant Enterococci spp. 15 The prolonged use of a lab coat and the greater number of coats used by a healthcare personnel are associated with higher number of isolates on the sleeve of the coat compared to other parts. 18 The white coat used by Jordanian healthcare personnel has long sleeves and two pockets at waist level, reaches the mid-thigh, and is made from a blend of polyester and cotton fabrics.
The hand hygiene of healthcare personnel plays an important role in microbial transmission. Touching items during or after the physical examination of patients without hand washing can act as a vehicle for the transmission of healthcare-associated infection to patients and objects. 19 Therefore, the aim of this study was to identify, describe, and explore the extent of bacterial contamination and the species of bacteria, particularly Escherichia coli (E. coli), on healthcare personnel attire. It also aimed to identify the hand hygiene practices of healthcare personnel that might reduce and control bacterial growth.

Methods
Design and setting A descriptive, cross-sectional design was used in the current study. Cross-sectional study is useful to assess and describe the health status of communities 20 in a relatively short period during data collection and less cost. Convenience sampling of 188 healthcare personnel, including physicians, nurses, and allied health (medicine, nursing and pharmacy) students, was used in our study. The current study was conducted in one of the main hospitals located in the northern part of Jordan.

Data and Sample Collection
Approval (Rf#20180170) from Jordan University of Science and Technology (JUST) was obtained to conduct this study. Each participant completed an anonymous, self-administered, and structured research questionnaire (socio-demographic data). Three swab samples were collected from three different locations (collar, mid-sleeve, and waist) on the lab coat or scrub used by the participants. Three milliliters of phosphate buffered saline was added to sterile cotton swabs before sample collection. An area of 10 by 10 cm of each three locations of the lab coat or scrub was swabbed. Then, the swabs were examined.

Total Plate Count and Sample Storage
Each swab sample was vortexed in a solution for 10 sec to release the bacteria into the solution. Serial dilution was performed and 100 µl from two dilutions were inoculated onto two nutrient agar plates to estimate the colony forming units for each sample. Colonies were counted after overnight incubation at 37°C. For stock keeping, 5 ml Tryptone soya broth (Oxoid, UK: CM0129) was added to each swab and incubated at 37°C for 48 h, followed by mixing 1 ml of the culture with 666 µl of 50% glycerol and storage at -80°C. Colonies were cultured on four different differential media to identify major bacterial pathogens as follows: Staphylococcus aureus was grown on Baird-Parker agar; Klebsiella pneumoniae and Escherichia coli were grown on MacConkey agar; Pseudomonas aeruginosa was grown on cetrimide agar; and Enterococcus faecalis was grown on bile esculin agar.

Statistical Analysis
The mean and standard deviation were used to describe continuous variables, whereas the frequencies and percentages were used for categorical variables. The multiple response analysis was used to describe the prevalence of the grown species across the three different lab coat locations (collar, midsleeve and waist), accounting for the dependency between those measured bacterial species. Cochran's Q chi-squared test was used to assess the correlation between the growth of speci c bacterial species across the locations, assuming that the three locations were repeated-measures for the same participant. The factor analysis for categorical variables was used to assess age, marital state, monthly income, number of family members that the healthcare personnel live with, and specialization. ANOVA test was used for the categorical variables of three or more levels. The generalized mixed linear model was used to identify the predictors of bacterial growth on healthcare personnel attire.
One hundred eighty-eight healthcare personnel and allied health students participated in the current study. The mean age of the participants was 28.10 (SD = 5.67) y, ranging from 19 to 43 y old. The distribution of the age groups of healthcare personnel was as follows: 52.7% were aged between 19 and 27 y; 37.2%, between 28 and 35 y, and 10.1%, between 36 and 43 y. More than half of the participants were women (57.4%). In addition, about 53.2% of the participants were never married. Of the participants, allied health students accounted for 31.4%; nurses, 36.7%; and physicians, 31.9%. The department-wise distribution of participants was as follows: surgical oors, 21.8%; medical oors, 33.5%; critical care units, 17.6%; outpatient clinics and auxiliary departments, 12.8%; and emergency department, 14.4% (Table 1).
With respect to clinical experience, 29.8% of participants had less than 1 y experience; 29.8%, between 2 and 5 y; and 40.4%, >5 y. As for economic status, 31.4% of the participants had household monthly income of less than 400 JOD. The mean household size or number of family members was 4.15 (SD = 2.3) members. Majority of the participants lived with their spouses and children (28.7%). Table 2 displays the handling behaviors and the hygiene and washing practices of uniforms or lab coats used by the healthcare personnel. Around half (50.5%) of the personnel preferred to wear white lab coats over scrubs for their daily work. The average number of uniforms or lab coats owned by the healthcare personnel was 2.1 (SD = 0.82). The mean number of washing times of healthcare personnel attire per month was 7.42 (SD = 6.86). Most of the participants (55.3%) washed their uniforms between 1 and 4 times per month and 41.5% of the participants washed their uniforms or lab coats once every three days.
Of the participants, 29.8% owned only one uniform or lab coat. The majority of the participants (53.2%) kept their uniforms or lab coats inside hospital lockers, whereas 46.8% kept theirs at home.
Around half (54.3%) of the participants owned the same uniform or lab coat for less than 1 y, while 45.7% owned theirs for more than 1 y. Only 13.3% of the healthcare personnel borrowed their uniforms or lab coats from their work peers; 36.7% were assigned or had contact to ten patients or fewer every day; 27.7% had contact to 10 to 15 patients every day; and most of the participants (35.6%) had contact to or cared for more than 15 patients every day. The majority (83%) mentioned that they wear uniforms or lab coats to comply with professional attire or dress code. In addition, 96.3% of the participants agreed that uniforms and white lab coats can be potential vehicles of pathogens to their patients (Table. 2). Table 3 displays the bacterial growth ndings yielded from screening healthcare personnel attire using surveillance swabs of three locations of each attire. Regardless of the location, presumptive growth types of bacteria were as follows: 77.7% of the healthcare personnel attire had E. faecalis; 31.2%, S. aureus; 28%, E. coli; and 5.1%, P. aeruginosa. The mean of number of colony forming units in the inoculated petri dishes was 9.81 (SD = 4.46). Pertaining to certain locations on healthcare personnel attire, the distribution of positive bacterial growth was as follows: the collar had mean colony forming units 3. The proportions of the species of bacteria across three different locations of healthcare personnel attire were compared using the Cochran's Q chi-squared test for signi cant differences. The analysis ndings showed that none of the four species (Enterococcus faecalis (p = 0.224), Staphylococcus aureus (p = 0.711), Escherichia coli (p = 0.057) and Pseudomonas aeruginosa (p = 0.197)) differed signi cantly between the collar, mid-sleeve, and waist (Table 4).
Moreover, the one-way ANOVA revealed no statistically signi cant mean differences in the number of recovered organism CFU across healthcare workers from various hospital units (p=0.080). However, the HCP attire from critical care units and emergency rooms showed a slightly higher mean recovered CFU ( Figure 2).
The generalized mixed linear model (Table 5) showed that the attire of healthcare personnel aged between 28 and 35 y exhibited signi cantly lower mean colony forming units compared to those of personnel aged ≥36 y (p = 0.021). The same model revealed that attire worn by male healthcare personnel exhibited signi cantly lower mean colony forming units compared to those worn by female personnel, accounting for other predictors in the analysis (beta = -0.112, p = 0.007). The participants with household monthly income less than 400 JOD exhibited signi cantly lower mean colony forming units compared to those with monthly income more than 800 JOD (p = 0.011), accounting for the other predictors. Allied health students' attire exhibited signi cantly lower mean colony forming units compared to physicians (p < 0.001). The attire of healthcare personnel who work in critical care units, showed signi cantly higher mean colony forming units than the attire of those who work in the emergency department (beta = 0.283, p < 0.001), accounting for other predictors in the model. The same analysis model revealed that healthcare personnel who live alone measured signi cantly higher mean attire colony forming units compared to those who live in shared residence (beta = 0.225, p < 0.001). healthcare personnel living with spouse, children, and parents showed signi cantly greater mean attire colony forming units compared to those living with friends/others (beta = 0.119, p = 0.020). The healthcare personnel who used white coats exhibited signi cantly lower mean attire colony forming units compared to Scrubs (beta = -0.124, p = 0.014). The attire of healthcare personnel who owned only one uniform had signi cantly lower mean attire colony forming units compared to those of personnel who owned three uniforms (beta = -0.174, p < 0.001).
In addition, the analysis model showed that the personnel who did not borrow their uniforms from their peers exhibited slightly lower mean attire colony forming units compared to those who borrowed their uniforms from other work peer (beta = -0.092, p = 0.068). The attire of personnel who prefer to carry their uniforms covered by anything other than bag had signi cantly higher mean attire colony forming units than those of personnel who carry their uniforms by hand without cover (beta = 0.121, p = 0.006). Moreover, the attire of personnel who carried their uniforms in a bag exhibited signi cantly lower mean attire colony forming units compared to those of personnel who carried their uniforms by hand and without cover or bag (beta = -0.100, p = 0.022).
Because of the nearly signi cant growth differences of Escherichiacoli (p=0.057) at different locations on lab coats from other types of bacteria (Table 4), we performed a bivariate analysis for the occurrence of Escherichiacoli in association with the HCPs' sociodemographic, professional, and attire hygiene practice levels (Table 6).
In the chi-square test, positive E.coli growth on HCP attire was associated with the female sex (p=0.032; Table 6) and old age (mean: 29.10 ± 4.35 vs. 27.92 ± 5.82 years; p=0.046). Furthermore, in the chi-square test for analysing the association of the likelihood of positive E. coli growth across HCWs of different age groups, those aged between 28 and 38 years showed signi cantly more growth of E. coli (p=0.035) compared to HCWs of other age groups.
The marital state of HCP did not correlate signi cantly with their likelihood of having positive E. coli growth on their attire (p=0.066; Table 6). However, the ever-married people were slightly more inclined to have positive E. coli growths on their attire. The chi-square test showed that students were signi cantly less likely to have positive E. coli growth compared to nurses and physicians (p<0.001). Interestingly, the HCP working in critical care had signi cantly less positive E. coli growth on their attire compared to those working in other departments (p=0.002). HCP working in surgical oors showed less positive E. coli growths, but those working in the emergency room or treating outpatients showed more positive E. coli growth on their attire.
In Figure 3, the odds of having E. coli growth on the attire has been depicted on the y-axis, and HCPs' working/training area has been depicted on the y-axis. It is clear that people working in emergency rooms and outpatient areas are the most susceptible, followed by those in medical and surgical oors and then those in critical care.
Moreover, the duration of experience of HCP correlated signi cantly with the probability of having positive E. coli growth on their attire (p=0.022). The chi-squared test showed that those with an experience of 2 to 5 years were more likely to have E. coli growth on their attire. The monthly income of HCP's households correlated signi cantly with their likelihood of having E. coli growth on their attire (p<0.001). People with an income of 600-800 JOD were signi cantly more inclined to have positive E. coli growth swabs (p<0.001) than the other HCP with different monthly income levels of the households. Nonetheless, a standardized socioeconomic index score was computed using the factor analysis procedure comprising the HCP's income, age, marital status, and living conditions. The mean socioeconomic index score differed signi cantly between people with and without E. coli growth on their attire. Those who had the growth showed a higher socioeconomic index score compared to those who did not (p=0.004; Table 6).
Moreover, HCP with family size of 4 to 6 members were signi cantly more likely to have E. coli growth on their hospital attire (p=0.012), according to the chi-squared test. However, the analysis results suggested that the residence type of HCP did not correlate signi cantly with their probability of positive E. coli growth on their hospital attire (p=0.406). HCP living in houses shared with their spouse, kids, and parents were slightly more likely to have positive E. coli samples on their hospital attires (Table 6). Table 6 shows the bivariate analysis results for the association between HCP's attire characteristics and hygienic practices with the likelihood of having E. coli-positive samples on various attire sites. The analysis outcomes suggested that the HCP using various types of uniforms did not differ with respect to their attire growing E. coli (p=0.550). Moreover, their rate of washing the attire did not correlate signi cantly with their likelihood of having positive E. coli samples on their attire sites (p=0.957). Furthermore, HCP's hygiene and attire characteristics and practices (namely the number of lab coats, washing frequency of lab coats, age of lab coats, borrowing lab coats from other workers, reasons of wearing lab coats, and the way they handle them besides the laundry service used for their attire) correlated with their likelihood of having E. coli samples on their attire (p>0.050 each). However, the chisquared test suggested that the HCP who store their attire in the hospital were signi cantly more inclined to show positive E. coli samples on various attire sites compared to those who stored their attire at home (p=0.002). Moreover, the analysis suggested that HCP who cared for 10 to 15 patients per day were signi cantly more likely to have positive E. coli samples compared to other HCP who deliver care to <10 patients per day or those who care for >15 patients per day (p<0.001). Storing the lab coat at the hospital correlated signi cantly with a higher incidence of positive E. coli samples on the attire (p=0.039). Likewise, the HCP who wore their uniforms outside hospital settings were found to be signi cantly more likely to have positive E. coli samples on their attire (p=0.005). Those who disagreed with the ability of hospital attire to transmit pathogens were slightly more likely to have positive E. coli samples on their attire (p=0.081), according to the chi-squared test of independence.

Discussion
In our study, Enterococcus faecalis was the most common species of bacteria found on healthcare personnel attire. Notwithstanding, Enterococcus usually does not cause health problems in healthy people, unless it has spread to other areas of the body which might cause life-threatening infections, 21 such as endocarditis, sepsis, urinary tract infections (UTIs), and meningitis. 22 E. faecalis is found in a wide variety of environments, including soil, water, and human and animal gastrointestinal tract. In our study, the source of E. faecalis contamination in healthcare personnel attire might be due to poor hospital surface disinfection, personal hygiene, or both. Moreover, E. faecalis is able to survive on dry surfaces for nearly a week to four months, depending on the strain as stated by Wendt and colleagues. 23 Whereas, Staphylococcus aureus was the most common contamination agent on lab coats in studies conducted in Moshi, Tanzania 24 and Mansoura, Egypt 25 , which contradict the results obtained in the current study. In our study, the presence of E. coli varied based on the location on the lab coat. In the literature, it has been associated with numerous diseases, such as in ammatory bowel disease 26 , Crohn's disease 27 , UTIs 28 , and pediatric acute diarrhea. 29 In the current study, the distribution of bacterial growth was highest on the mid-sleeve area healthcare personnel attire, in which mid-sleeves are close to the sides of uniforms. This is in concordance with what was found in the study conducted by Banu and colleagues 30 . They reported that the sides of the coats were the most highly contaminated areas compared to the collar and pockets. The mid-sleeves might be contaminated when resting arms on nurse stations or reception desks while documenting activities. Mid-sleeve areas are the prominent parts of the body that might touch surrounding environment while moving and ambulating. E. coli was found to be the highest in the collar area of the lab coat in the current study. This might be due to transmission from the perineal area to the collar because of a bad washing habit and changing of lab coats in bathrooms.
To the best of our knowledge, the current study only reveals the correlation between the age of healthcare personnel and the bacterial contamination on their attire. This study indicated that healthcare personnel aged ≥ 36 y had higher bacterial contamination in their attire compared to those in other age groups. Younger healthcare personnel possibly tend to change and clean their attire more frequently. In contrast, E. coli was found to be the highest in the attire of HCP aged between 28 and 36 years, which might be due to the frequent use of bathrooms among younger HCWs.
The attire of female healthcare personnel had higher bacterial contamination than those of male personnel, which is different from what was reported by Akanbi and colleagues. 15 Their study revealed that lab coats of male resident doctors were more contaminated compared to female resident doctors. The justi cation of the results in our study is that most of allied health students, including female allied health students, have only one place, which is bathrooms, to change their dresses due to the lack of specialized dressing rooms. This was supported by what was observed during the data collection stage as reported by Dir and colleagues 31 . In their study, they observed that many people attended the bathroom while wearing their lab coats. Similarly, female HCP attires were highly contaminated with E. coli compared to their male counterparts in this study. UTIs caused by E. coli is more common in women 32 , which raises concerns about female HCP being prone to UTI from their attires.
Allied health students' uniforms, in our study, had lower contamination compared to those of physicians, which could be due to the higher adherence to university o cial code of tidy and clean lab coats/scrubs by the students than by the physicians. This result might be unanticipated compared to what was reported by Muhadi and colleagues, 33 who demonstrated higher contamination in allied health students' attire and that allied health students use lab coats outside the hospital premises, resulting in the higher contamination levels of their lab coats, particularly E. coli growth on nurses' attires and those working on medical oors. The type of HCP working department in our study also showed a signi cant role in causing higher contamination of the attire among HCP working in critical care units compared to the emergency department. In concordance with the role of the department in bacterial contamination, the surgery department was found to predispose its HCP to the highest degree of contamination, as reported by Akanbi and colleagues. 15 In the current study, white lab coats were more contaminated compared to scrubs, which in turn, raised questions about the importance of wearing white lab coats in settings where patients received healthcare. This was also recommended by Qaday and colleagues, 24 in which lab coats should not be recommended to use anymore. Although Banu and colleagues reported that the possession of two or more white lab coats reduced contamination compared to the possession of only one, 30 our study results showed the opposite-that those who possess two lab coats were more contaminated. This might be due to failure to keep track of which lab coat needs to be cleaning. This result could be also justi ed that HCP might clean their only one lab coat frequently compared to who possess several lab coats. Hygienic practices of lab coats/scrubs might be considered more important than the number of uniforms that HCP possess in terms of bacterial contamination on uniforms.
Borrowed lab coats/scrubs from others in the current study had high contamination, which necessitates the compulsory policy of possession at least two coat/scrub for each HCP as well as yearly purchase of lab coats/scrubs and banning the borrowing lab coats from each other as this item is strictly personal. 34 In addition, due to the borrowing of these attire, the owners of lab coats/scrubs cannot keep track of where these labs coats were used and when they should be washed. Despite the recommendation that lab coats/scrubs should be stored within hospital premises 35 and washed regularly because of continuous bacterial contaminations within a few hours of usage, the current study revealed that E. coli grew on attires stored in hospital lockers more than those sored at homes. A washing facility on the hospital premises is a must to keep the lab coats clean and cut the chain of contamination of nosocomial pathogens, even after newly laundered uniforms. 36 It was shocking that the positive growth of E. coli on attires were associated with the number of cared/examined patients per day (more than ten patients) by HCP in our study. Such contamination of E. coli predisposes patients to E. coli bacteremia, which in turn leads to adverse health outcomes, such as increased length of stay and cost of those patients 37 as well as advanced colorectal neoplasia 38 . As a result, infection control starting with combating contamination with E. coli and its strains should be structured and adhered to by HCP as part of their efforts in eradicating nosocomial infections [39][40][41] . Subsequently, further steps should be antimicrobial susceptibility testing to contribute signi cantly in the treatment of nosocomial infections for patients in the healthcare departments 42 .

Conclusions
The use of white lab coats develops more concerns pertinent to its role in transmission healthcareassociated infection compared to use of scrubs. HCP are in urgent need to recognize the risk factors and behaviors that contribute to healthcare-associated infection through use of lab coats, not merely hygienic handling of lab coats, but awareness of places or hospital premises where lab coats should be used. HCP, including physicians, nurses, and allied health students, should be cautious on how and where lab coats should be brought and worn.  Note: Medical students are medicine, nursing and pharmacy students.    Figure 1 The mean of bacterial growth CFUs measured from three bodily sites of the HCPs' attire The association between HCP working unit and their attire total recovered Mean-Log(CFU) counts