Assessment of microbial bioburden on portable medical equipment in a hospital setting

Abstract

Objectives:

Although routine disinfection of portable medical equipment is required in most hospitals, frontline staff may not be able to disinfect portable medical equipment at a rate that adequately maintains low bioburden on high-use equipment. This study quantified bioburden over an extended time period for two types of portable medical equipment, workstations on wheels and vitals machines, across three hospital wards.

Methods:

Bioburden was quantified via press plate samples taken from high touch surfaces on 10 workstations on wheels and 5 vitals machines on each of 3 medical surgical units. The samples were taken at three timepoints each day over a 4-week period, with random rotation of timepoints and portable medical equipment, such that frontline staff were not aware at which timepoint their portable medical equipment would be sampled. The mean bioburden from the different locations and portable medical equipment was estimated and compared with Bayesian multilevel negative binomial regression models.

Results:

Model estimated mean colony counts (95% credible interval) were 14.4 (7.7–26.7) for vitals machines and 29.2 (16.1–51.1) for workstations on wheels. For the workstations on wheel, colony counts were lower on the mouse, 0.22 (0.16–0.29), tray, 0.29 (0.22, 0.38), and keyboard, 0.43 (0.32–0.55), when compared to the arm, as assessed by incident rate ratios.

Conclusions:

Although routine disinfection is required, bioburden is still present across portable medical equipment on a variety of surfaces. The difference in bioburden levels among surfaces likely reflects differences in touch patterns for the different portable medical equipment and surfaces on the portable medical equipment. Although the association of portable medical equipment bioburden to healthcare-associated infection transmission was not assessed, this study provides evidence for the potential of portable medical equipment as a vector for healthcare-associated infection transmission despite hospital disinfection requirements.

Introduction

Healthcare-associated infections (HAIs) are a significant cause of morbidity and mortality among hospitalized patients. Despite evidence that infection prevention practices can reduce HAIs,^1,2 adherence to infection prevention protocols remains sub-optimal.^3,4 Common infection prevention strategies include hand hygiene and environmental cleaning and disinfection to mitigate transmission of pathogens. The ability to monitor adherence to standard infection prevention protocols is an important component of linking outcomes to practice.^5–7 Adherence to protocols may be challenged by increased workload and complexity of demands on nursing and ancillary staff for patient care, such as those experienced during the Coronavirus Diseases-19 (COVID-19) pandemic. Limitations on ability to adequately disinfect contaminated surfaces threaten the inferred effectiveness of established protocols demonstrated to prevent HAIs and may be important contributors to recent estimates by the Centers for Disease Control and Prevention (CDC) that 1 in 31 hospital patients contracts a HAI.

Previous studies have primarily focused on environmental contamination in the patient room (high touch surfaces such as bedrails, tray tables, call buttons, etc.) that may contribute to HAIs.^8–10 Several pathogens also have the ability to survive on portable medical equipment (PME) that moves through the patient care environment and between patient rooms. ¹¹ Transmission to patient or surfaces in a patient room can occur via “direct” interaction if the equipment or fomite contacts the patient or component of the room environment, and via “indirect” interaction when personnel touch the equipment or fomite and then touch the patient. ¹² PME are designated in the noncritical medical device category for infection risk according to the Spaulding Classification Scheme. ¹³ Disinfection for this device category is determined by individual hospital and infection prevention control groups. Devices that come in direct contact with patients are recommended to be disinfected after each use. Workstations on wheels (WOWs) or Vitals Machines (VMs) in our hospital policy are to be manually disinfected once a day and as needed (if used in contact isolation rooms or presence of bodily fluid). Typical end users for WOWs and VMs include nursing staff and occasionally medical officers. An updated information for disinfectants is available to the end users from a grid for the type of disinfectant and contact times needed for use on each PME in our hospital. The purpose of this study was to determine if current disinfection practices were effectively managing bioburden on WOWs and VMs in our hospital wards.

Methods:This study was conducted at the Central Texas Veterans Health Care System (CTVHCS) hospital, Temple, TX in the acute care units and approved by the research committees. This study was a prospective single hospital observational study and exempted by Institutional Review Board (IRB) (Federal Wide Assurance number FWA00001125) at the Central Texas Veterans Health Care System (CTVHCS), Temple, Texas. WOWs and VMs in use on these units were sampled by the research team for aerobic bacterial colonies. The sampling was performed using contact Rodac plates (25 cm², Hardy Diagnostics, Santa Maria, CA). The details of the sampling method, incubation, and plate count are described previously. ⁸ Data was collected longitudinally between April and May, 2019 for 10 WOWs and 5 VMs on each of 3 units in the hospital over a 4-week period, 4 consecutive days per week (Monday to Thursday). Sampling was performed at three timepoints each day—9 am, noon, and 3 pm. On any given day, samples were collected on the unit at only one time, but the collection times were rotated throughout the days of the study so that samples were collected for all three timepoints for all 3 units. In addition, multiple locations on each PME were sampled. On WOWs, the keyboard, mouse, arm, and tray were sampled. On the VMs, the upper right and left, and bottom right and left, of the VM machine were sampled (Figure 1(a) and (b). Since only one sample per PME was taken at a time, these locations were randomly assigned to attempt an unbiased and balanced sample of sites from different units, PME, days, and hours of the day.

Figure 1.

(a) A workstation on wheels (left) showing the areas that are sampled for the study in yellow. (b) A vitals machine (right) showing the areas that are sampled for the study.

Statistical analysis

Three Bayesian multilevel negative binomial models were used to estimate the mean bioburden on WOWs and VMs and make comparisons: bioburden on WOWs versus VMs, and comparison of bioburden on locations of both types of PME. Varying intercepts were used to account for the potential correlation of bioburden within the same unit and PME device. In addition, for the model on all data comparing PME type, the shape (over-dispersion) parameter for the negative binomial distribution was predicted by PME type, relaxing the assumption that both WOWs and VMs would have similar distributions of counts. The models were run in the brms package in R and plots made utilizing ggplot2. Results are reported as estimated means (plate colony counts) with 95% Uncertainty Intervals (UIs; 95% quantiles of the posterior credible interval), and as incidence rate ratios (IRRs) with 95% UI, the ratio of rate of counts per plate.

Results

There were 480 planned samples for WOWs and 240 planned samples for VMs. However, 28 samples from WOWs and 30 samples from VMs were missing, due to the PME not being available on the unit at the sampling time.

The estimated mean colony counts and 95% UI for press plates taken from WOWs was 29.2 (16.1–51.1) counts and on the VMs was 14.4 (7.7–26.7) counts. The IRR for WOWs compared to VMs was 2.1 (1.6–2.7); thus, the model estimates show around twice as many counts for the WOW compared to the VM.

Among sampling locations for WOWs, the estimated mean colony counts were 58.8 (35.2–91.2) for the arm, 24.7 (14.5–37.2) for the keyboard, 17.0 (10.0–26.4) for the tray, and 12.5 (7.4–19.3) for the mouse. The IRRs for keyboard compared to arm was 0.43 (0.32–0.55), for tray compared to arm was 0.29 (0.22–0.38), and for mouse compared to arm was 0.22 (0.16–0.29) (Figure 2(a)).

Figure 2.

(a) Colony counts taken from press plates on each day of the study for each of the three hospital units. Sampling locations on the WOW are denoted by shape and color. A small amount of horizontal jitter has been added to avoid overlap of points. (b) Colony counts taken from press plates on each day of the study for each of the three hospital units. Sampling locations on the VMs are denoted by shape and color. A small amount of horizontal jitter has been added to avoid overlap of points.

Among sampling locations for VMs, the estimated mean colony counts were 24.4 (11.9–44.8) for the left bottom, 11.4 (5.7–21.1) for the left top, 14.9 (7.4–28.2) for the right bottom, and 9.7 (4.7–17.6) for the right top. The IRRs for left top compared to left bottom was 0.48 (0.30–0.73), for right bottom compared to left bottom was 0.63 (0.38–0.99), and for right top compared to left bottom was 0.41 (0.25–0.62; Figure 2(b)).

Discussion

PME present in the patient care environment is often found to be contaminated with pathogens. Our previous sequence analysis study highlighted the patterns and sequence of contact events of PME including WOWs and healthcare workers and patients in a hospital setting. ¹² Our present study identified contamination levels of different PME and the parts/regions of the PME that are more highly contaminated. The results from our study indicate that WOWs in this small study had approximately twice the bioburden compared to VMs. This is not surprising given that WOWs are in constant use by the healthcare worker, whereas VMs in our facility are typically used intermittently and placed in a holding area between uses or left in patient rooms for the duration of their stay. Among WOWs, the arm had greater bioburden levels than the keyboard, tray, or mouse. This may be due to the practice of moving the WOW about via the arm. Although the keyboard and the mouse experience a large amount of contact, these surfaces may receive more focused cleaning. It is also possible that the differences in bioburden levels among these locations is due to the sampling process rather than the result of the bioburden accumulation process, as the keyboard and mouse may be more difficult to sample with a press plate. Among VMs, the bottom left of the machine showed the highest levels of bioburden. This may reflect holding the machine with the left hand while pressing the buttons with the right hand, or the result may simply be due to noise and a smaller sample size.

Researchers have used adenosine triphosphate bioluminescence assays, aerobic cultures, and fluorescent markers to assess the cleanliness of PME disinfected by staff members between each patient use.^14–16 Havill, et al. reported that portable equipment was often not adequately cleaned according to written protocols following use. ¹⁴ In a previous study we used a novel disinfection tracking system that can be attached to any PME and that continuously records any disinfection events. Findings from that study included a pattern of disinfection events at a frequency higher than self-reported compliance to the standard hospital policy of a minimum of once every 24 h and improvement in disinfection when the instantaneous end-user feedback was turned on.^6,7 Our current study indicates that, despite adherence to our hospital policy, PME remains contaminated. Implications of this finding may indicate more frequent disinfection of PME, especially high touch areas may be advisable. Recommendations of disinfection frequency will need to be informed by a better understanding of causes of variations in surface contamination, patterns of hand hygiene compared to patterns of work behaviors, and the relationship between types, locations, and level of surface contamination with HAIs as well as availability of dedicated staff or nursing time to keep up with increased frequency.

Our study has several limitations as discussed here. This is a small study with limited number of equipment that were monitored in a single hospital. Several studies have reported recovery of HAI-causing pathogens such as methicillin resistant Staphylococcus aureus (MRSA), multi-drug resistant Acinetobacter, vancomycin-resistant Enterococcus, etc. from hospital surfaces as summarized in a systematic review. ² In a recent separate study, we have reported the presence of MRSA on WOWs. ¹⁷ Another limitation is that any species identification for the aerobic bacterial colonies (ABCs) to determine recovery of commensals, multi-drug resistant organisms (MDROs) or other HAI-causing pathogens on PME was not performed for this study. Earlier cluster (c)-RCTs (5 out of 14) conducted on reducing MDROs and or HAIs, in a hospital environment indicated that either interventions such as use of cleaning and disinfecting agents (UV, copper, and bleach) or educational programs may be effective. ¹⁸ Finally, our study does not directly link contamination of equipment to transmission of pathogens.

Conclusions

The contamination pattern of PMEs (WOWs and VMs) were observed longitudinally and reflected higher bioburden for WOWs than VMs. Our study also suggests that despite adherence to standard hospital policy, the PMEs remain contaminated with a need for more frequent cleaning and disinfection.